Search – talk-data.com

Title & Speakers	Event
April 24, 2025 - AI, Machine Learning and Computer Vision Meetup 2025-04-24 · 17:00 This is a virtual event. Register for the Zoom Towards a Multimodal AI Agent that Can See, Talk and Act The development of multimodal AI agents marks a pivotal step toward creating systems capable of understanding, reasoning, and interacting with the world in human-like ways. Building such agents requires models that not only comprehend multi-sensory observations but also act adaptively to achieve goals within their environments. In this talk, I will present my research journey toward this grand goal across three key dimensions. First, I will explore how to bridge the gap between core vision understanding and multimodal learning through unified frameworks at various granularities. Next, I will discuss connecting vision-language models with large language models (LLMs) to create intelligent conversational systems. Finally, I will delve into recent advancements that extend multimodal LLMs into vision-language-action models, forming the foundation for general-purpose robotics policies. To conclude, I will highlight ongoing efforts to develop agentic systems that integrate perception with action, enabling them to not only understand observations but also take meaningful actions in a single system. Together, these lead to an aspiration of building the next generation of multimodal AI agents capable of seeing, talking, and acting across diverse scenarios in both digital and physical worlds. About the Speaker Jianwei Yang is a Principal Researcher at Microsoft Research (MSR), Redmond. His research focuses on the intersection of vision and multimodal learning, with an emphasis on bridging core vision tasks with language, building general-purpose and promptable multimodal models, and enabling these models to take meaningful actions in both virtual and physical environments. ConceptAttention: Interpreting the Representations of Diffusion Transformers Recently, diffusion transformers have taken over as the state-of-the-art model class for both image and video generation. However, similar to many existing deep learning architectures, their high-dimensional hidden representations are difficult to understand and interpret. This lack of interpretability is a barrier to their controllability and safe deployment. We introduce ConceptAttention, an approach to interpreting the representations of diffusion transformers. Our method allows users to create rich saliency maps depicting the location and intensity of textual concepts. Our approach exposes how a diffusion model “sees” a generated image and notably requires no additional training. ConceptAttention improves upon widely used approaches like cross attention maps for isolating the location of visual concepts and even generalizes to real world (not just generated) images and video generation models! Our work serves to improve the community’s understanding of how diffusion models represent data and has numerous potential applications, like image editing. About the Speaker Alec Helbling is a PhD student at Georgia Tech. His research focuses on improving the interpretability and controllability of generative models, particularly for image generation. His research is more application focused, and he has have interned at a variety of industrial research labs like Adobe Firefly, IBM Research, and NASA Jet Propulsion Lab. He also has a passion for creating explanatory videos of interesting machine learning and mathematical concepts. RelationField: Relate Anything in Radiance Fields Neural radiance fields recently emerged as a 3D scene representation extended by distilling open-vocabulary features from vision-language models. Current methods focus on object-centric tasks, leaving semantic relationships largely unexplored. We propose RelationField, the first method extracting inter-object relationships directly from neural radiance fields using pairs of rays for implicit relationship queries. RelationField distills relationship knowledge from multi-modal LLMs. Evaluated on open-vocabulary 3D scene graph generation and relationship-guided instance segmentation, RelationField achieves state-of-the-art performance. About the Speaker Sebastian Koch is a PhD student at Ulm University and Bosch Center for Artificial Intelligence. He is supervised by Timo Ropinski from Ulm University. His main research interest lies at the intersection of computer vision and robotics. The goal of his PhD is to develop 3D scene representations of the real world that are valuable for robots to navigate and solve tasks within their environment RGB-X Model Development: Exploring Four Channel ML Workflows Machine Learning is rapidly becoming multimodal. With many models in Computer Vision expanding to areas like vision and 3D, one area that has also quietly been advancing rapidly is RGB-X data, such as infrared, depth, or normals. In this talk we will cover some of the leading models in this exploding field of Visual AI and show some best practices on how to work with these complex data formats! About the Speaker Daniel Gural is a seasoned Machine Learning Evangelist with a strong passion for empowering Data Scientists and ML Engineers to unlock the full potential of their data. Currently serving as a valuable member of Voxel51, he takes a leading role in efforts to bridge the gap between practitioners and the necessary tools, enabling them to achieve exceptional outcomes. Daniel’s extensive experience in teaching and developing within the ML field has fueled his commitment to democratizing high-quality AI workflows for a wider audience.	April 24, 2025 - AI, Machine Learning and Computer Vision Meetup
April 24, 2025 - AI, Machine Learning and Computer Vision Meetup 2025-04-24 · 17:00 This is a virtual event. Register for the Zoom Towards a Multimodal AI Agent that Can See, Talk and Act The development of multimodal AI agents marks a pivotal step toward creating systems capable of understanding, reasoning, and interacting with the world in human-like ways. Building such agents requires models that not only comprehend multi-sensory observations but also act adaptively to achieve goals within their environments. In this talk, I will present my research journey toward this grand goal across three key dimensions. First, I will explore how to bridge the gap between core vision understanding and multimodal learning through unified frameworks at various granularities. Next, I will discuss connecting vision-language models with large language models (LLMs) to create intelligent conversational systems. Finally, I will delve into recent advancements that extend multimodal LLMs into vision-language-action models, forming the foundation for general-purpose robotics policies. To conclude, I will highlight ongoing efforts to develop agentic systems that integrate perception with action, enabling them to not only understand observations but also take meaningful actions in a single system. Together, these lead to an aspiration of building the next generation of multimodal AI agents capable of seeing, talking, and acting across diverse scenarios in both digital and physical worlds. About the Speaker Jianwei Yang is a Principal Researcher at Microsoft Research (MSR), Redmond. His research focuses on the intersection of vision and multimodal learning, with an emphasis on bridging core vision tasks with language, building general-purpose and promptable multimodal models, and enabling these models to take meaningful actions in both virtual and physical environments. ConceptAttention: Interpreting the Representations of Diffusion Transformers Recently, diffusion transformers have taken over as the state-of-the-art model class for both image and video generation. However, similar to many existing deep learning architectures, their high-dimensional hidden representations are difficult to understand and interpret. This lack of interpretability is a barrier to their controllability and safe deployment. We introduce ConceptAttention, an approach to interpreting the representations of diffusion transformers. Our method allows users to create rich saliency maps depicting the location and intensity of textual concepts. Our approach exposes how a diffusion model “sees” a generated image and notably requires no additional training. ConceptAttention improves upon widely used approaches like cross attention maps for isolating the location of visual concepts and even generalizes to real world (not just generated) images and video generation models! Our work serves to improve the community’s understanding of how diffusion models represent data and has numerous potential applications, like image editing. About the Speaker Alec Helbling is a PhD student at Georgia Tech. His research focuses on improving the interpretability and controllability of generative models, particularly for image generation. His research is more application focused, and he has have interned at a variety of industrial research labs like Adobe Firefly, IBM Research, and NASA Jet Propulsion Lab. He also has a passion for creating explanatory videos of interesting machine learning and mathematical concepts. RelationField: Relate Anything in Radiance Fields Neural radiance fields recently emerged as a 3D scene representation extended by distilling open-vocabulary features from vision-language models. Current methods focus on object-centric tasks, leaving semantic relationships largely unexplored. We propose RelationField, the first method extracting inter-object relationships directly from neural radiance fields using pairs of rays for implicit relationship queries. RelationField distills relationship knowledge from multi-modal LLMs. Evaluated on open-vocabulary 3D scene graph generation and relationship-guided instance segmentation, RelationField achieves state-of-the-art performance. About the Speaker Sebastian Koch is a PhD student at Ulm University and Bosch Center for Artificial Intelligence. He is supervised by Timo Ropinski from Ulm University. His main research interest lies at the intersection of computer vision and robotics. The goal of his PhD is to develop 3D scene representations of the real world that are valuable for robots to navigate and solve tasks within their environment RGB-X Model Development: Exploring Four Channel ML Workflows Machine Learning is rapidly becoming multimodal. With many models in Computer Vision expanding to areas like vision and 3D, one area that has also quietly been advancing rapidly is RGB-X data, such as infrared, depth, or normals. In this talk we will cover some of the leading models in this exploding field of Visual AI and show some best practices on how to work with these complex data formats! About the Speaker Daniel Gural is a seasoned Machine Learning Evangelist with a strong passion for empowering Data Scientists and ML Engineers to unlock the full potential of their data. Currently serving as a valuable member of Voxel51, he takes a leading role in efforts to bridge the gap between practitioners and the necessary tools, enabling them to achieve exceptional outcomes. Daniel’s extensive experience in teaching and developing within the ML field has fueled his commitment to democratizing high-quality AI workflows for a wider audience.	April 24, 2025 - AI, Machine Learning and Computer Vision Meetup
April 24, 2025 - AI, Machine Learning and Computer Vision Meetup 2025-04-24 · 17:00 This is a virtual event. Register for the Zoom Towards a Multimodal AI Agent that Can See, Talk and Act The development of multimodal AI agents marks a pivotal step toward creating systems capable of understanding, reasoning, and interacting with the world in human-like ways. Building such agents requires models that not only comprehend multi-sensory observations but also act adaptively to achieve goals within their environments. In this talk, I will present my research journey toward this grand goal across three key dimensions. First, I will explore how to bridge the gap between core vision understanding and multimodal learning through unified frameworks at various granularities. Next, I will discuss connecting vision-language models with large language models (LLMs) to create intelligent conversational systems. Finally, I will delve into recent advancements that extend multimodal LLMs into vision-language-action models, forming the foundation for general-purpose robotics policies. To conclude, I will highlight ongoing efforts to develop agentic systems that integrate perception with action, enabling them to not only understand observations but also take meaningful actions in a single system. Together, these lead to an aspiration of building the next generation of multimodal AI agents capable of seeing, talking, and acting across diverse scenarios in both digital and physical worlds. About the Speaker Jianwei Yang is a Principal Researcher at Microsoft Research (MSR), Redmond. His research focuses on the intersection of vision and multimodal learning, with an emphasis on bridging core vision tasks with language, building general-purpose and promptable multimodal models, and enabling these models to take meaningful actions in both virtual and physical environments. ConceptAttention: Interpreting the Representations of Diffusion Transformers Recently, diffusion transformers have taken over as the state-of-the-art model class for both image and video generation. However, similar to many existing deep learning architectures, their high-dimensional hidden representations are difficult to understand and interpret. This lack of interpretability is a barrier to their controllability and safe deployment. We introduce ConceptAttention, an approach to interpreting the representations of diffusion transformers. Our method allows users to create rich saliency maps depicting the location and intensity of textual concepts. Our approach exposes how a diffusion model “sees” a generated image and notably requires no additional training. ConceptAttention improves upon widely used approaches like cross attention maps for isolating the location of visual concepts and even generalizes to real world (not just generated) images and video generation models! Our work serves to improve the community’s understanding of how diffusion models represent data and has numerous potential applications, like image editing. About the Speaker Alec Helbling is a PhD student at Georgia Tech. His research focuses on improving the interpretability and controllability of generative models, particularly for image generation. His research is more application focused, and he has have interned at a variety of industrial research labs like Adobe Firefly, IBM Research, and NASA Jet Propulsion Lab. He also has a passion for creating explanatory videos of interesting machine learning and mathematical concepts. RelationField: Relate Anything in Radiance Fields Neural radiance fields recently emerged as a 3D scene representation extended by distilling open-vocabulary features from vision-language models. Current methods focus on object-centric tasks, leaving semantic relationships largely unexplored. We propose RelationField, the first method extracting inter-object relationships directly from neural radiance fields using pairs of rays for implicit relationship queries. RelationField distills relationship knowledge from multi-modal LLMs. Evaluated on open-vocabulary 3D scene graph generation and relationship-guided instance segmentation, RelationField achieves state-of-the-art performance. About the Speaker Sebastian Koch is a PhD student at Ulm University and Bosch Center for Artificial Intelligence. He is supervised by Timo Ropinski from Ulm University. His main research interest lies at the intersection of computer vision and robotics. The goal of his PhD is to develop 3D scene representations of the real world that are valuable for robots to navigate and solve tasks within their environment. RGB-X Model Development: Exploring Four Channel ML Workflows Machine Learning is rapidly becoming multimodal. With many models in Computer Vision expanding to areas like vision and 3D, one area that has also quietly been advancing rapidly is RGB-X data, such as infrared, depth, or normals. In this talk we will cover some of the leading models in this exploding field of Visual AI and show some best practices on how to work with these complex data formats! About the Speaker Daniel Gural is a seasoned Machine Learning Evangelist with a strong passion for empowering Data Scientists and ML Engineers to unlock the full potential of their data. Currently serving as a valuable member of Voxel51, he takes a leading role in efforts to bridge the gap between practitioners and the necessary tools, enabling them to achieve exceptional outcomes. Daniel’s extensive experience in teaching and developing within the ML field has fueled his commitment to democratizing high-quality AI workflows for a wider audience.	April 24, 2025 - AI, Machine Learning and Computer Vision Meetup
April 24, 2025 - AI, Machine Learning and Computer Vision Meetup 2025-04-24 · 17:00 This is a virtual event. Register for the Zoom Towards a Multimodal AI Agent that Can See, Talk and Act The development of multimodal AI agents marks a pivotal step toward creating systems capable of understanding, reasoning, and interacting with the world in human-like ways. Building such agents requires models that not only comprehend multi-sensory observations but also act adaptively to achieve goals within their environments. In this talk, I will present my research journey toward this grand goal across three key dimensions. First, I will explore how to bridge the gap between core vision understanding and multimodal learning through unified frameworks at various granularities. Next, I will discuss connecting vision-language models with large language models (LLMs) to create intelligent conversational systems. Finally, I will delve into recent advancements that extend multimodal LLMs into vision-language-action models, forming the foundation for general-purpose robotics policies. To conclude, I will highlight ongoing efforts to develop agentic systems that integrate perception with action, enabling them to not only understand observations but also take meaningful actions in a single system. Together, these lead to an aspiration of building the next generation of multimodal AI agents capable of seeing, talking, and acting across diverse scenarios in both digital and physical worlds. About the Speaker Jianwei Yang is a Principal Researcher at Microsoft Research (MSR), Redmond. His research focuses on the intersection of vision and multimodal learning, with an emphasis on bridging core vision tasks with language, building general-purpose and promptable multimodal models, and enabling these models to take meaningful actions in both virtual and physical environments. ConceptAttention: Interpreting the Representations of Diffusion Transformers Recently, diffusion transformers have taken over as the state-of-the-art model class for both image and video generation. However, similar to many existing deep learning architectures, their high-dimensional hidden representations are difficult to understand and interpret. This lack of interpretability is a barrier to their controllability and safe deployment. We introduce ConceptAttention, an approach to interpreting the representations of diffusion transformers. Our method allows users to create rich saliency maps depicting the location and intensity of textual concepts. Our approach exposes how a diffusion model “sees” a generated image and notably requires no additional training. ConceptAttention improves upon widely used approaches like cross attention maps for isolating the location of visual concepts and even generalizes to real world (not just generated) images and video generation models! Our work serves to improve the community’s understanding of how diffusion models represent data and has numerous potential applications, like image editing. About the Speaker Alec Helbling is a PhD student at Georgia Tech. His research focuses on improving the interpretability and controllability of generative models, particularly for image generation. His research is more application focused, and he has have interned at a variety of industrial research labs like Adobe Firefly, IBM Research, and NASA Jet Propulsion Lab. He also has a passion for creating explanatory videos of interesting machine learning and mathematical concepts. RelationField: Relate Anything in Radiance Fields Neural radiance fields recently emerged as a 3D scene representation extended by distilling open-vocabulary features from vision-language models. Current methods focus on object-centric tasks, leaving semantic relationships largely unexplored. We propose RelationField, the first method extracting inter-object relationships directly from neural radiance fields using pairs of rays for implicit relationship queries. RelationField distills relationship knowledge from multi-modal LLMs. Evaluated on open-vocabulary 3D scene graph generation and relationship-guided instance segmentation, RelationField achieves state-of-the-art performance. About the Speaker Sebastian Koch is a PhD student at Ulm University and Bosch Center for Artificial Intelligence. He is supervised by Timo Ropinski from Ulm University. His main research interest lies at the intersection of computer vision and robotics. The goal of his PhD is to develop 3D scene representations of the real world that are valuable for robots to navigate and solve tasks within their environment RGB-X Model Development: Exploring Four Channel ML Workflows Machine Learning is rapidly becoming multimodal. With many models in Computer Vision expanding to areas like vision and 3D, one area that has also quietly been advancing rapidly is RGB-X data, such as infrared, depth, or normals. In this talk we will cover some of the leading models in this exploding field of Visual AI and show some best practices on how to work with these complex data formats! About the Speaker Daniel Gural is a seasoned Machine Learning Evangelist with a strong passion for empowering Data Scientists and ML Engineers to unlock the full potential of their data. Currently serving as a valuable member of Voxel51, he takes a leading role in efforts to bridge the gap between practitioners and the necessary tools, enabling them to achieve exceptional outcomes. Daniel’s extensive experience in teaching and developing within the ML field has fueled his commitment to democratizing high-quality AI workflows for a wider audience.	April 24, 2025 - AI, Machine Learning and Computer Vision Meetup
April 24, 2025 - AI, Machine Learning and Computer Vision Meetup 2025-04-24 · 17:00 This is a virtual event. Register for the Zoom Towards a Multimodal AI Agent that Can See, Talk and Act The development of multimodal AI agents marks a pivotal step toward creating systems capable of understanding, reasoning, and interacting with the world in human-like ways. Building such agents requires models that not only comprehend multi-sensory observations but also act adaptively to achieve goals within their environments. In this talk, I will present my research journey toward this grand goal across three key dimensions. First, I will explore how to bridge the gap between core vision understanding and multimodal learning through unified frameworks at various granularities. Next, I will discuss connecting vision-language models with large language models (LLMs) to create intelligent conversational systems. Finally, I will delve into recent advancements that extend multimodal LLMs into vision-language-action models, forming the foundation for general-purpose robotics policies. To conclude, I will highlight ongoing efforts to develop agentic systems that integrate perception with action, enabling them to not only understand observations but also take meaningful actions in a single system. Together, these lead to an aspiration of building the next generation of multimodal AI agents capable of seeing, talking, and acting across diverse scenarios in both digital and physical worlds. About the Speaker Jianwei Yang is a Principal Researcher at Microsoft Research (MSR), Redmond. His research focuses on the intersection of vision and multimodal learning, with an emphasis on bridging core vision tasks with language, building general-purpose and promptable multimodal models, and enabling these models to take meaningful actions in both virtual and physical environments. ConceptAttention: Interpreting the Representations of Diffusion Transformers Recently, diffusion transformers have taken over as the state-of-the-art model class for both image and video generation. However, similar to many existing deep learning architectures, their high-dimensional hidden representations are difficult to understand and interpret. This lack of interpretability is a barrier to their controllability and safe deployment. We introduce ConceptAttention, an approach to interpreting the representations of diffusion transformers. Our method allows users to create rich saliency maps depicting the location and intensity of textual concepts. Our approach exposes how a diffusion model “sees” a generated image and notably requires no additional training. ConceptAttention improves upon widely used approaches like cross attention maps for isolating the location of visual concepts and even generalizes to real world (not just generated) images and video generation models! Our work serves to improve the community’s understanding of how diffusion models represent data and has numerous potential applications, like image editing. About the Speaker Alec Helbling is a PhD student at Georgia Tech. His research focuses on improving the interpretability and controllability of generative models, particularly for image generation. His research is more application focused, and he has have interned at a variety of industrial research labs like Adobe Firefly, IBM Research, and NASA Jet Propulsion Lab. He also has a passion for creating explanatory videos of interesting machine learning and mathematical concepts. RelationField: Relate Anything in Radiance Fields Neural radiance fields recently emerged as a 3D scene representation extended by distilling open-vocabulary features from vision-language models. Current methods focus on object-centric tasks, leaving semantic relationships largely unexplored. We propose RelationField, the first method extracting inter-object relationships directly from neural radiance fields using pairs of rays for implicit relationship queries. RelationField distills relationship knowledge from multi-modal LLMs. Evaluated on open-vocabulary 3D scene graph generation and relationship-guided instance segmentation, RelationField achieves state-of-the-art performance. About the Speaker Sebastian Koch is a PhD student at Ulm University and Bosch Center for Artificial Intelligence. He is supervised by Timo Ropinski from Ulm University. His main research interest lies at the intersection of computer vision and robotics. The goal of his PhD is to develop 3D scene representations of the real world that are valuable for robots to navigate and solve tasks within their environment RGB-X Model Development: Exploring Four Channel ML Workflows Machine Learning is rapidly becoming multimodal. With many models in Computer Vision expanding to areas like vision and 3D, one area that has also quietly been advancing rapidly is RGB-X data, such as infrared, depth, or normals. In this talk we will cover some of the leading models in this exploding field of Visual AI and show some best practices on how to work with these complex data formats! About the Speaker Daniel Gural is a seasoned Machine Learning Evangelist with a strong passion for empowering Data Scientists and ML Engineers to unlock the full potential of their data. Currently serving as a valuable member of Voxel51, he takes a leading role in efforts to bridge the gap between practitioners and the necessary tools, enabling them to achieve exceptional outcomes. Daniel’s extensive experience in teaching and developing within the ML field has fueled his commitment to democratizing high-quality AI workflows for a wider audience.	April 24, 2025 - AI, Machine Learning and Computer Vision Meetup
April 24, 2025 - AI, Machine Learning and Computer Vision Meetup 2025-04-24 · 17:00 This is a virtual event. Register for the Zoom Towards a Multimodal AI Agent that Can See, Talk and Act The development of multimodal AI agents marks a pivotal step toward creating systems capable of understanding, reasoning, and interacting with the world in human-like ways. Building such agents requires models that not only comprehend multi-sensory observations but also act adaptively to achieve goals within their environments. In this talk, I will present my research journey toward this grand goal across three key dimensions. First, I will explore how to bridge the gap between core vision understanding and multimodal learning through unified frameworks at various granularities. Next, I will discuss connecting vision-language models with large language models (LLMs) to create intelligent conversational systems. Finally, I will delve into recent advancements that extend multimodal LLMs into vision-language-action models, forming the foundation for general-purpose robotics policies. To conclude, I will highlight ongoing efforts to develop agentic systems that integrate perception with action, enabling them to not only understand observations but also take meaningful actions in a single system. Together, these lead to an aspiration of building the next generation of multimodal AI agents capable of seeing, talking, and acting across diverse scenarios in both digital and physical worlds. About the Speaker Jianwei Yang is a Principal Researcher at Microsoft Research (MSR), Redmond. His research focuses on the intersection of vision and multimodal learning, with an emphasis on bridging core vision tasks with language, building general-purpose and promptable multimodal models, and enabling these models to take meaningful actions in both virtual and physical environments. ConceptAttention: Interpreting the Representations of Diffusion Transformers Recently, diffusion transformers have taken over as the state-of-the-art model class for both image and video generation. However, similar to many existing deep learning architectures, their high-dimensional hidden representations are difficult to understand and interpret. This lack of interpretability is a barrier to their controllability and safe deployment. We introduce ConceptAttention, an approach to interpreting the representations of diffusion transformers. Our method allows users to create rich saliency maps depicting the location and intensity of textual concepts. Our approach exposes how a diffusion model “sees” a generated image and notably requires no additional training. ConceptAttention improves upon widely used approaches like cross attention maps for isolating the location of visual concepts and even generalizes to real world (not just generated) images and video generation models! Our work serves to improve the community’s understanding of how diffusion models represent data and has numerous potential applications, like image editing. About the Speaker Alec Helbling is a PhD student at Georgia Tech. His research focuses on improving the interpretability and controllability of generative models, particularly for image generation. His research is more application focused, and he has have interned at a variety of industrial research labs like Adobe Firefly, IBM Research, and NASA Jet Propulsion Lab. He also has a passion for creating explanatory videos of interesting machine learning and mathematical concepts. RelationField: Relate Anything in Radiance Fields Neural radiance fields recently emerged as a 3D scene representation extended by distilling open-vocabulary features from vision-language models. Current methods focus on object-centric tasks, leaving semantic relationships largely unexplored. We propose RelationField, the first method extracting inter-object relationships directly from neural radiance fields using pairs of rays for implicit relationship queries. RelationField distills relationship knowledge from multi-modal LLMs. Evaluated on open-vocabulary 3D scene graph generation and relationship-guided instance segmentation, RelationField achieves state-of-the-art performance. About the Speaker Sebastian Koch is a PhD student at Ulm University and Bosch Center for Artificial Intelligence. He is supervised by Timo Ropinski from Ulm University. His main research interest lies at the intersection of computer vision and robotics. The goal of his PhD is to develop 3D scene representations of the real world that are valuable for robots to navigate and solve tasks within their environment RGB-X Model Development: Exploring Four Channel ML Workflows Machine Learning is rapidly becoming multimodal. With many models in Computer Vision expanding to areas like vision and 3D, one area that has also quietly been advancing rapidly is RGB-X data, such as infrared, depth, or normals. In this talk we will cover some of the leading models in this exploding field of Visual AI and show some best practices on how to work with these complex data formats! About the Speaker Daniel Gural is a seasoned Machine Learning Evangelist with a strong passion for empowering Data Scientists and ML Engineers to unlock the full potential of their data. Currently serving as a valuable member of Voxel51, he takes a leading role in efforts to bridge the gap between practitioners and the necessary tools, enabling them to achieve exceptional outcomes. Daniel’s extensive experience in teaching and developing within the ML field has fueled his commitment to democratizing high-quality AI workflows for a wider audience.	April 24, 2025 - AI, Machine Learning and Computer Vision Meetup
April 24, 2025 - AI, Machine Learning and Computer Vision Meetup 2025-04-24 · 17:00 This is a virtual event. Register for the Zoom Towards a Multimodal AI Agent that Can See, Talk and Act The development of multimodal AI agents marks a pivotal step toward creating systems capable of understanding, reasoning, and interacting with the world in human-like ways. Building such agents requires models that not only comprehend multi-sensory observations but also act adaptively to achieve goals within their environments. In this talk, I will present my research journey toward this grand goal across three key dimensions. First, I will explore how to bridge the gap between core vision understanding and multimodal learning through unified frameworks at various granularities. Next, I will discuss connecting vision-language models with large language models (LLMs) to create intelligent conversational systems. Finally, I will delve into recent advancements that extend multimodal LLMs into vision-language-action models, forming the foundation for general-purpose robotics policies. To conclude, I will highlight ongoing efforts to develop agentic systems that integrate perception with action, enabling them to not only understand observations but also take meaningful actions in a single system. Together, these lead to an aspiration of building the next generation of multimodal AI agents capable of seeing, talking, and acting across diverse scenarios in both digital and physical worlds. About the Speaker Jianwei Yang is a Principal Researcher at Microsoft Research (MSR), Redmond. His research focuses on the intersection of vision and multimodal learning, with an emphasis on bridging core vision tasks with language, building general-purpose and promptable multimodal models, and enabling these models to take meaningful actions in both virtual and physical environments. ConceptAttention: Interpreting the Representations of Diffusion Transformers Recently, diffusion transformers have taken over as the state-of-the-art model class for both image and video generation. However, similar to many existing deep learning architectures, their high-dimensional hidden representations are difficult to understand and interpret. This lack of interpretability is a barrier to their controllability and safe deployment. We introduce ConceptAttention, an approach to interpreting the representations of diffusion transformers. Our method allows users to create rich saliency maps depicting the location and intensity of textual concepts. Our approach exposes how a diffusion model “sees” a generated image and notably requires no additional training. ConceptAttention improves upon widely used approaches like cross attention maps for isolating the location of visual concepts and even generalizes to real world (not just generated) images and video generation models! Our work serves to improve the community’s understanding of how diffusion models represent data and has numerous potential applications, like image editing. About the Speaker Alec Helbling is a PhD student at Georgia Tech. His research focuses on improving the interpretability and controllability of generative models, particularly for image generation. His research is more application focused, and he has have interned at a variety of industrial research labs like Adobe Firefly, IBM Research, and NASA Jet Propulsion Lab. He also has a passion for creating explanatory videos of interesting machine learning and mathematical concepts. RelationField: Relate Anything in Radiance Fields Neural radiance fields recently emerged as a 3D scene representation extended by distilling open-vocabulary features from vision-language models. Current methods focus on object-centric tasks, leaving semantic relationships largely unexplored. We propose RelationField, the first method extracting inter-object relationships directly from neural radiance fields using pairs of rays for implicit relationship queries. RelationField distills relationship knowledge from multi-modal LLMs. Evaluated on open-vocabulary 3D scene graph generation and relationship-guided instance segmentation, RelationField achieves state-of-the-art performance. About the Speaker Sebastian Koch is a PhD student at Ulm University and Bosch Center for Artificial Intelligence. He is supervised by Timo Ropinski from Ulm University. His main research interest lies at the intersection of computer vision and robotics. The goal of his PhD is to develop 3D scene representations of the real world that are valuable for robots to navigate and solve tasks within their environment. RGB-X Model Development: Exploring Four Channel ML Workflows Machine Learning is rapidly becoming multimodal. With many models in Computer Vision expanding to areas like vision and 3D, one area that has also quietly been advancing rapidly is RGB-X data, such as infrared, depth, or normals. In this talk we will cover some of the leading models in this exploding field of Visual AI and show some best practices on how to work with these complex data formats! About the Speaker Daniel Gural is a seasoned Machine Learning Evangelist with a strong passion for empowering Data Scientists and ML Engineers to unlock the full potential of their data. Currently serving as a valuable member of Voxel51, he takes a leading role in efforts to bridge the gap between practitioners and the necessary tools, enabling them to achieve exceptional outcomes. Daniel’s extensive experience in teaching and developing within the ML field has fueled his commitment to democratizing high-quality AI workflows for a wider audience.	April 24, 2025 - AI, Machine Learning and Computer Vision Meetup
Generative AI Meetup (Paris) 2024-04-30 · 16:00 Important RSVP HERE (Due to room capacity and venue security, it is required to pre-register at the link for admission) Description: Welcome to the AI meetup in Paris. In collaboration with Roosh Circle, a European tech community for entrepreneurs, engineers, VC representatives, and business experts. This meetup is devoted to Generative AI — the technology that is evolving rapidly and has great potential across diverse sectors. Let’s meet offline to discuss the power of GenAI across industries and strategies for integrating it into business. Agenda: - 6:00-6:30pm: Gathering & Welcome Drinks - 6:30-6:40pm: Intro Word from Roosh Circle - 6:40-8:10pm: Tech talks and Q&A - 8:10-9:30pm: Networking & Drinks . Tech Talk: Agile in the AI Era Speaker: Yassine AQEJJAJ (Devoteam) Abstract: In this talk, Yassine will dive deeper into the transformative impact of AI on Agile methodologies, focusing on software development. Tech Talk: Scaling RAG: From Concept to Production Speaker: Louis GERMAIN (Ipsos Synthesio) Abstract: In this talk, Louise will share his experience with RAG evaluation and the challenges of scaling the feature with cost efficiency, stability, and speed. Tech Talk: Advancing the LLM Deployment with Secure European Cloud Speaker: Frédéric BARDOLLE (Scaleway) Abstract: In this talk, Frédéric will share tactics for initiating secure deployment of AI projects, and shed light on how a resilient and sustainable cloud ecosystem can enhance LLMs utilization. For whom: – AI/DL/ML Engineers with 3+ years of experience – ML researchers and professionals – Individuals with a solid foundation in GenAI – Product managers in AI-focused companies When: April 30, 6 – 9 pm. Registration deadline: April 29 Venue:** Liberte Living-Lab, 9 Rue d'Alexandrie, Atrium Space Join Roosh Circle Community on Discord and LinkedIn Event hosts: Roosh Circle and AICamp. Let's meet and improve the world with the power of GenAI	Generative AI Meetup (Paris)
Event AI Meetup (April): Generative AI, LLMs and ML 2024-04-09
Tech Talk: MultiModal Learning and Synthetic AI 2024-04-09 · 21:30 This talk will be a Deep Dive into Multi-Modal AI Models, the powerful AI systems that are behind the functioning of applications such GPT Vision, DALL-E, and even Sora. We will go through the core theory of how intelligence from different sources of data (text, AI, and Vision) are combined to together in order help build capabilities like real-time Image/Video analysis, image hyper segmentation, and image to text extraction, and more. gpt vision dall-e sora
Tech Talk: Build Lightweight LLMs Applications with Open Source Tools 2024-04-09 · 21:30 ChatGPT is awesome, but developing with its API comes at a cost. Fortunately, there are open-source alternatives like Google Gemini, Streamlit, and Python APIs that can fetch prompt results using an API key. In this presentation, I'll explore how to create a lightweight, self-service end-to-end LLMs application using prompt engineering and fine-tuning based on user requests. Additionally, I'll demonstrate how to build a food suggestion application based on ingredients or food names. google gemini streamlit python apis prompt engineering fine-tuning llms

AI Meetup (April): Generative AI, LLMs and ML 2024-04-09 · 21:30 Important RSVP here (Due to room capacity and building security, you must pre-register at the link for admission) Description: Welcome to our in-person AI meetup in New York. Join us for deep dive tech talks on AI, GenAI, LLMs and ML, hands-on workshops, food/drink, networking with speakers and fellow developers. Tech Talk: Build Lightweight LLMs Applications with Open Source Tools Speaker: Amogh Kokari (Con Edison) Abstract: ChatGPT is awesome, but developing with its API comes at a cost. Fortunately, there are open-source alternatives like Google Gemini, Streamlit, and Python APIs that can fetch prompt results using an API key. In this presentation, I'll explore how to create a lightweight, self-service end-to-end LLMs application using prompt engineering and fine-tuning based on user requests. Additionally, I'll demonstrate how to build a food suggestion application based on ingredients or food names. Tech Talk: MultiModal Learning and Synthetic AI Speaker: Shafik Quoraishee (The New York Times) Abstract: This talk will be a Deep Dive into Multi-Modal AI Models, the powerful AI systems that are behind the functioning of applications such GPT Vision, DALL-E, and even Sora. We will go through the core theory of how intelligence from different sources of data (text, AI, and Vision) are combined to together in order help build capabilities like real-time Image/Video analysis, image hyper segmentation, and image to text extraction, and more. Speakers/Topics: Stay tuned as we are updating speakers and schedules. If you have a keen interest in speaking to our community, we invite you to submit topics for consideration: Submit Topics Sponsors: We are actively seeking sponsors to support our community. Whether it is by offering venue spaces, providing food/drink, or cash sponsor. Sponsors will not only speak at the meetups, receive prominent recognition, but also gain exposure to our extensive membership base of 20,000+ AI developers in New York or 350K+ worldwide. Community on Slack/Discord** Event chat: chat and connect with speakers and attendees Sharing blogs, events, job openings, projects collaborations Join Slack/Discord (scroll down to the bottom) *	AI Meetup (April): Generative AI, LLMs and ML

April 24, 2025 - AI, Machine Learning and Computer Vision Meetup 2025-04-24 · 17:00

This is a virtual event.

Towards a Multimodal AI Agent that Can See, Talk and Act

The development of multimodal AI agents marks a pivotal step toward creating systems capable of understanding, reasoning, and interacting with the world in human-like ways. Building such agents requires models that not only comprehend multi-sensory observations but also act adaptively to achieve goals within their environments. In this talk, I will present my research journey toward this grand goal across three key dimensions.

First, I will explore how to bridge the gap between core vision understanding and multimodal learning through unified frameworks at various granularities. Next, I will discuss connecting vision-language models with large language models (LLMs) to create intelligent conversational systems. Finally, I will delve into recent advancements that extend multimodal LLMs into vision-language-action models, forming the foundation for general-purpose robotics policies. To conclude, I will highlight ongoing efforts to develop agentic systems that integrate perception with action, enabling them to not only understand observations but also take meaningful actions in a single system.

Together, these lead to an aspiration of building the next generation of multimodal AI agents capable of seeing, talking, and acting across diverse scenarios in both digital and physical worlds.

About the Speaker

Jianwei Yang is a Principal Researcher at Microsoft Research (MSR), Redmond. His research focuses on the intersection of vision and multimodal learning, with an emphasis on bridging core vision tasks with language, building general-purpose and promptable multimodal models, and enabling these models to take meaningful actions in both virtual and physical environments.

ConceptAttention: Interpreting the Representations of Diffusion Transformers

Recently, diffusion transformers have taken over as the state-of-the-art model class for both image and video generation. However, similar to many existing deep learning architectures, their high-dimensional hidden representations are difficult to understand and interpret. This lack of interpretability is a barrier to their controllability and safe deployment.

We introduce ConceptAttention, an approach to interpreting the representations of diffusion transformers. Our method allows users to create rich saliency maps depicting the location and intensity of textual concepts. Our approach exposes how a diffusion model “sees” a generated image and notably requires no additional training. ConceptAttention improves upon widely used approaches like cross attention maps for isolating the location of visual concepts and even generalizes to real world (not just generated) images and video generation models!

Our work serves to improve the community’s understanding of how diffusion models represent data and has numerous potential applications, like image editing.

About the Speaker

Alec Helbling is a PhD student at Georgia Tech. His research focuses on improving the interpretability and controllability of generative models, particularly for image generation. His research is more application focused, and he has have interned at a variety of industrial research labs like Adobe Firefly, IBM Research, and NASA Jet Propulsion Lab. He also has a passion for creating explanatory videos of interesting machine learning and mathematical concepts.

RelationField: Relate Anything in Radiance Fields

Neural radiance fields recently emerged as a 3D scene representation extended by distilling open-vocabulary features from vision-language models. Current methods focus on object-centric tasks, leaving semantic relationships largely unexplored. We propose RelationField, the first method extracting inter-object relationships directly from neural radiance fields using pairs of rays for implicit relationship queries. RelationField distills relationship knowledge from multi-modal LLMs. Evaluated on open-vocabulary 3D scene graph generation and relationship-guided instance segmentation, RelationField achieves state-of-the-art performance.

About the Speaker

Sebastian Koch is a PhD student at Ulm University and Bosch Center for Artificial Intelligence. He is supervised by Timo Ropinski from Ulm University. His main research interest lies at the intersection of computer vision and robotics. The goal of his PhD is to develop 3D scene representations of the real world that are valuable for robots to navigate and solve tasks within their environment

RGB-X Model Development: Exploring Four Channel ML Workflows

Machine Learning is rapidly becoming multimodal. With many models in Computer Vision expanding to areas like vision and 3D, one area that has also quietly been advancing rapidly is RGB-X data, such as infrared, depth, or normals. In this talk we will cover some of the leading models in this exploding field of Visual AI and show some best practices on how to work with these complex data formats!

About the Speaker

Daniel Gural is a seasoned Machine Learning Evangelist with a strong passion for empowering Data Scientists and ML Engineers to unlock the full potential of their data. Currently serving as a valuable member of Voxel51, he takes a leading role in efforts to bridge the gap between practitioners and the necessary tools, enabling them to achieve exceptional outcomes. Daniel’s extensive experience in teaching and developing within the ML field has fueled his commitment to democratizing high-quality AI workflows for a wider audience.

April 24, 2025 - AI, Machine Learning and Computer Vision Meetup

April 24, 2025 - AI, Machine Learning and Computer Vision Meetup 2025-04-24 · 17:00

This is a virtual event.

Register for the Zoom

Towards a Multimodal AI Agent that Can See, Talk and Act

The development of multimodal AI agents marks a pivotal step toward creating systems capable of understanding, reasoning, and interacting with the world in human-like ways. Building such agents requires models that not only comprehend multi-sensory observations but also act adaptively to achieve goals within their environments. In this talk, I will present my research journey toward this grand goal across three key dimensions.

First, I will explore how to bridge the gap between core vision understanding and multimodal learning through unified frameworks at various granularities. Next, I will discuss connecting vision-language models with large language models (LLMs) to create intelligent conversational systems. Finally, I will delve into recent advancements that extend multimodal LLMs into vision-language-action models, forming the foundation for general-purpose robotics policies. To conclude, I will highlight ongoing efforts to develop agentic systems that integrate perception with action, enabling them to not only understand observations but also take meaningful actions in a single system.

Together, these lead to an aspiration of building the next generation of multimodal AI agents capable of seeing, talking, and acting across diverse scenarios in both digital and physical worlds.

About the Speaker

Jianwei Yang is a Principal Researcher at Microsoft Research (MSR), Redmond. His research focuses on the intersection of vision and multimodal learning, with an emphasis on bridging core vision tasks with language, building general-purpose and promptable multimodal models, and enabling these models to take meaningful actions in both virtual and physical environments.

ConceptAttention: Interpreting the Representations of Diffusion Transformers

Recently, diffusion transformers have taken over as the state-of-the-art model class for both image and video generation. However, similar to many existing deep learning architectures, their high-dimensional hidden representations are difficult to understand and interpret. This lack of interpretability is a barrier to their controllability and safe deployment.

We introduce ConceptAttention, an approach to interpreting the representations of diffusion transformers. Our method allows users to create rich saliency maps depicting the location and intensity of textual concepts. Our approach exposes how a diffusion model “sees” a generated image and notably requires no additional training. ConceptAttention improves upon widely used approaches like cross attention maps for isolating the location of visual concepts and even generalizes to real world (not just generated) images and video generation models!

Our work serves to improve the community’s understanding of how diffusion models represent data and has numerous potential applications, like image editing.

About the Speaker

Alec Helbling is a PhD student at Georgia Tech. His research focuses on improving the interpretability and controllability of generative models, particularly for image generation. His research is more application focused, and he has have interned at a variety of industrial research labs like Adobe Firefly, IBM Research, and NASA Jet Propulsion Lab. He also has a passion for creating explanatory videos of interesting machine learning and mathematical concepts.

RelationField: Relate Anything in Radiance Fields

Neural radiance fields recently emerged as a 3D scene representation extended by distilling open-vocabulary features from vision-language models. Current methods focus on object-centric tasks, leaving semantic relationships largely unexplored. We propose RelationField, the first method extracting inter-object relationships directly from neural radiance fields using pairs of rays for implicit relationship queries. RelationField distills relationship knowledge from multi-modal LLMs. Evaluated on open-vocabulary 3D scene graph generation and relationship-guided instance segmentation, RelationField achieves state-of-the-art performance.

About the Speaker

Sebastian Koch is a PhD student at Ulm University and Bosch Center for Artificial Intelligence. He is supervised by Timo Ropinski from Ulm University. His main research interest lies at the intersection of computer vision and robotics. The goal of his PhD is to develop 3D scene representations of the real world that are valuable for robots to navigate and solve tasks within their environment

RGB-X Model Development: Exploring Four Channel ML Workflows

Machine Learning is rapidly becoming multimodal. With many models in Computer Vision expanding to areas like vision and 3D, one area that has also quietly been advancing rapidly is RGB-X data, such as infrared, depth, or normals. In this talk we will cover some of the leading models in this exploding field of Visual AI and show some best practices on how to work with these complex data formats!

About the Speaker

Daniel Gural is a seasoned Machine Learning Evangelist with a strong passion for empowering Data Scientists and ML Engineers to unlock the full potential of their data. Currently serving as a valuable member of Voxel51, he takes a leading role in efforts to bridge the gap between practitioners and the necessary tools, enabling them to achieve exceptional outcomes. Daniel’s extensive experience in teaching and developing within the ML field has fueled his commitment to democratizing high-quality AI workflows for a wider audience.

April 24, 2025 - AI, Machine Learning and Computer Vision Meetup

April 24, 2025 - AI, Machine Learning and Computer Vision Meetup 2025-04-24 · 17:00

This is a virtual event.

Register for the Zoom

Towards a Multimodal AI Agent that Can See, Talk and Act

The development of multimodal AI agents marks a pivotal step toward creating systems capable of understanding, reasoning, and interacting with the world in human-like ways. Building such agents requires models that not only comprehend multi-sensory observations but also act adaptively to achieve goals within their environments. In this talk, I will present my research journey toward this grand goal across three key dimensions.

First, I will explore how to bridge the gap between core vision understanding and multimodal learning through unified frameworks at various granularities. Next, I will discuss connecting vision-language models with large language models (LLMs) to create intelligent conversational systems. Finally, I will delve into recent advancements that extend multimodal LLMs into vision-language-action models, forming the foundation for general-purpose robotics policies. To conclude, I will highlight ongoing efforts to develop agentic systems that integrate perception with action, enabling them to not only understand observations but also take meaningful actions in a single system.

Together, these lead to an aspiration of building the next generation of multimodal AI agents capable of seeing, talking, and acting across diverse scenarios in both digital and physical worlds.

About the Speaker

Jianwei Yang is a Principal Researcher at Microsoft Research (MSR), Redmond. His research focuses on the intersection of vision and multimodal learning, with an emphasis on bridging core vision tasks with language, building general-purpose and promptable multimodal models, and enabling these models to take meaningful actions in both virtual and physical environments.

ConceptAttention: Interpreting the Representations of Diffusion Transformers

Recently, diffusion transformers have taken over as the state-of-the-art model class for both image and video generation. However, similar to many existing deep learning architectures, their high-dimensional hidden representations are difficult to understand and interpret. This lack of interpretability is a barrier to their controllability and safe deployment.

We introduce ConceptAttention, an approach to interpreting the representations of diffusion transformers. Our method allows users to create rich saliency maps depicting the location and intensity of textual concepts. Our approach exposes how a diffusion model “sees” a generated image and notably requires no additional training. ConceptAttention improves upon widely used approaches like cross attention maps for isolating the location of visual concepts and even generalizes to real world (not just generated) images and video generation models!

Our work serves to improve the community’s understanding of how diffusion models represent data and has numerous potential applications, like image editing.

About the Speaker

Alec Helbling is a PhD student at Georgia Tech. His research focuses on improving the interpretability and controllability of generative models, particularly for image generation. His research is more application focused, and he has have interned at a variety of industrial research labs like Adobe Firefly, IBM Research, and NASA Jet Propulsion Lab. He also has a passion for creating explanatory videos of interesting machine learning and mathematical concepts.

RelationField: Relate Anything in Radiance Fields

Neural radiance fields recently emerged as a 3D scene representation extended by distilling open-vocabulary features from vision-language models. Current methods focus on object-centric tasks, leaving semantic relationships largely unexplored. We propose RelationField, the first method extracting inter-object relationships directly from neural radiance fields using pairs of rays for implicit relationship queries. RelationField distills relationship knowledge from multi-modal LLMs. Evaluated on open-vocabulary 3D scene graph generation and relationship-guided instance segmentation, RelationField achieves state-of-the-art performance.

About the Speaker

Sebastian Koch is a PhD student at Ulm University and Bosch Center for Artificial Intelligence. He is supervised by Timo Ropinski from Ulm University. His main research interest lies at the intersection of computer vision and robotics. The goal of his PhD is to develop 3D scene representations of the real world that are valuable for robots to navigate and solve tasks within their environment.

RGB-X Model Development: Exploring Four Channel ML Workflows

Machine Learning is rapidly becoming multimodal. With many models in Computer Vision expanding to areas like vision and 3D, one area that has also quietly been advancing rapidly is RGB-X data, such as infrared, depth, or normals. In this talk we will cover some of the leading models in this exploding field of Visual AI and show some best practices on how to work with these complex data formats!

About the Speaker

Daniel Gural is a seasoned Machine Learning Evangelist with a strong passion for empowering Data Scientists and ML Engineers to unlock the full potential of their data. Currently serving as a valuable member of Voxel51, he takes a leading role in efforts to bridge the gap between practitioners and the necessary tools, enabling them to achieve exceptional outcomes. Daniel’s extensive experience in teaching and developing within the ML field has fueled his commitment to democratizing high-quality AI workflows for a wider audience.

April 24, 2025 - AI, Machine Learning and Computer Vision Meetup

April 24, 2025 - AI, Machine Learning and Computer Vision Meetup 2025-04-24 · 17:00

This is a virtual event.

Register for the Zoom

Towards a Multimodal AI Agent that Can See, Talk and Act

The development of multimodal AI agents marks a pivotal step toward creating systems capable of understanding, reasoning, and interacting with the world in human-like ways. Building such agents requires models that not only comprehend multi-sensory observations but also act adaptively to achieve goals within their environments. In this talk, I will present my research journey toward this grand goal across three key dimensions.

First, I will explore how to bridge the gap between core vision understanding and multimodal learning through unified frameworks at various granularities. Next, I will discuss connecting vision-language models with large language models (LLMs) to create intelligent conversational systems. Finally, I will delve into recent advancements that extend multimodal LLMs into vision-language-action models, forming the foundation for general-purpose robotics policies. To conclude, I will highlight ongoing efforts to develop agentic systems that integrate perception with action, enabling them to not only understand observations but also take meaningful actions in a single system.

Together, these lead to an aspiration of building the next generation of multimodal AI agents capable of seeing, talking, and acting across diverse scenarios in both digital and physical worlds.

About the Speaker

Jianwei Yang is a Principal Researcher at Microsoft Research (MSR), Redmond. His research focuses on the intersection of vision and multimodal learning, with an emphasis on bridging core vision tasks with language, building general-purpose and promptable multimodal models, and enabling these models to take meaningful actions in both virtual and physical environments.

ConceptAttention: Interpreting the Representations of Diffusion Transformers

Recently, diffusion transformers have taken over as the state-of-the-art model class for both image and video generation. However, similar to many existing deep learning architectures, their high-dimensional hidden representations are difficult to understand and interpret. This lack of interpretability is a barrier to their controllability and safe deployment.

We introduce ConceptAttention, an approach to interpreting the representations of diffusion transformers. Our method allows users to create rich saliency maps depicting the location and intensity of textual concepts. Our approach exposes how a diffusion model “sees” a generated image and notably requires no additional training. ConceptAttention improves upon widely used approaches like cross attention maps for isolating the location of visual concepts and even generalizes to real world (not just generated) images and video generation models!

Our work serves to improve the community’s understanding of how diffusion models represent data and has numerous potential applications, like image editing.

About the Speaker

Alec Helbling is a PhD student at Georgia Tech. His research focuses on improving the interpretability and controllability of generative models, particularly for image generation. His research is more application focused, and he has have interned at a variety of industrial research labs like Adobe Firefly, IBM Research, and NASA Jet Propulsion Lab. He also has a passion for creating explanatory videos of interesting machine learning and mathematical concepts.

RelationField: Relate Anything in Radiance Fields

Neural radiance fields recently emerged as a 3D scene representation extended by distilling open-vocabulary features from vision-language models. Current methods focus on object-centric tasks, leaving semantic relationships largely unexplored. We propose RelationField, the first method extracting inter-object relationships directly from neural radiance fields using pairs of rays for implicit relationship queries. RelationField distills relationship knowledge from multi-modal LLMs. Evaluated on open-vocabulary 3D scene graph generation and relationship-guided instance segmentation, RelationField achieves state-of-the-art performance.

About the Speaker

Sebastian Koch is a PhD student at Ulm University and Bosch Center for Artificial Intelligence. He is supervised by Timo Ropinski from Ulm University. His main research interest lies at the intersection of computer vision and robotics. The goal of his PhD is to develop 3D scene representations of the real world that are valuable for robots to navigate and solve tasks within their environment

RGB-X Model Development: Exploring Four Channel ML Workflows

Machine Learning is rapidly becoming multimodal. With many models in Computer Vision expanding to areas like vision and 3D, one area that has also quietly been advancing rapidly is RGB-X data, such as infrared, depth, or normals. In this talk we will cover some of the leading models in this exploding field of Visual AI and show some best practices on how to work with these complex data formats!

About the Speaker

Daniel Gural is a seasoned Machine Learning Evangelist with a strong passion for empowering Data Scientists and ML Engineers to unlock the full potential of their data. Currently serving as a valuable member of Voxel51, he takes a leading role in efforts to bridge the gap between practitioners and the necessary tools, enabling them to achieve exceptional outcomes. Daniel’s extensive experience in teaching and developing within the ML field has fueled his commitment to democratizing high-quality AI workflows for a wider audience.

April 24, 2025 - AI, Machine Learning and Computer Vision Meetup

April 24, 2025 - AI, Machine Learning and Computer Vision Meetup 2025-04-24 · 17:00

This is a virtual event.

Register for the Zoom

Towards a Multimodal AI Agent that Can See, Talk and Act

The development of multimodal AI agents marks a pivotal step toward creating systems capable of understanding, reasoning, and interacting with the world in human-like ways. Building such agents requires models that not only comprehend multi-sensory observations but also act adaptively to achieve goals within their environments. In this talk, I will present my research journey toward this grand goal across three key dimensions.

First, I will explore how to bridge the gap between core vision understanding and multimodal learning through unified frameworks at various granularities. Next, I will discuss connecting vision-language models with large language models (LLMs) to create intelligent conversational systems. Finally, I will delve into recent advancements that extend multimodal LLMs into vision-language-action models, forming the foundation for general-purpose robotics policies. To conclude, I will highlight ongoing efforts to develop agentic systems that integrate perception with action, enabling them to not only understand observations but also take meaningful actions in a single system.

Together, these lead to an aspiration of building the next generation of multimodal AI agents capable of seeing, talking, and acting across diverse scenarios in both digital and physical worlds.

About the Speaker

Jianwei Yang is a Principal Researcher at Microsoft Research (MSR), Redmond. His research focuses on the intersection of vision and multimodal learning, with an emphasis on bridging core vision tasks with language, building general-purpose and promptable multimodal models, and enabling these models to take meaningful actions in both virtual and physical environments.

ConceptAttention: Interpreting the Representations of Diffusion Transformers

Recently, diffusion transformers have taken over as the state-of-the-art model class for both image and video generation. However, similar to many existing deep learning architectures, their high-dimensional hidden representations are difficult to understand and interpret. This lack of interpretability is a barrier to their controllability and safe deployment.

We introduce ConceptAttention, an approach to interpreting the representations of diffusion transformers. Our method allows users to create rich saliency maps depicting the location and intensity of textual concepts. Our approach exposes how a diffusion model “sees” a generated image and notably requires no additional training. ConceptAttention improves upon widely used approaches like cross attention maps for isolating the location of visual concepts and even generalizes to real world (not just generated) images and video generation models!

Our work serves to improve the community’s understanding of how diffusion models represent data and has numerous potential applications, like image editing.

About the Speaker

Alec Helbling is a PhD student at Georgia Tech. His research focuses on improving the interpretability and controllability of generative models, particularly for image generation. His research is more application focused, and he has have interned at a variety of industrial research labs like Adobe Firefly, IBM Research, and NASA Jet Propulsion Lab. He also has a passion for creating explanatory videos of interesting machine learning and mathematical concepts.

RelationField: Relate Anything in Radiance Fields

Neural radiance fields recently emerged as a 3D scene representation extended by distilling open-vocabulary features from vision-language models. Current methods focus on object-centric tasks, leaving semantic relationships largely unexplored. We propose RelationField, the first method extracting inter-object relationships directly from neural radiance fields using pairs of rays for implicit relationship queries. RelationField distills relationship knowledge from multi-modal LLMs. Evaluated on open-vocabulary 3D scene graph generation and relationship-guided instance segmentation, RelationField achieves state-of-the-art performance.

About the Speaker

Sebastian Koch is a PhD student at Ulm University and Bosch Center for Artificial Intelligence. He is supervised by Timo Ropinski from Ulm University. His main research interest lies at the intersection of computer vision and robotics. The goal of his PhD is to develop 3D scene representations of the real world that are valuable for robots to navigate and solve tasks within their environment

RGB-X Model Development: Exploring Four Channel ML Workflows

Machine Learning is rapidly becoming multimodal. With many models in Computer Vision expanding to areas like vision and 3D, one area that has also quietly been advancing rapidly is RGB-X data, such as infrared, depth, or normals. In this talk we will cover some of the leading models in this exploding field of Visual AI and show some best practices on how to work with these complex data formats!

About the Speaker

Daniel Gural is a seasoned Machine Learning Evangelist with a strong passion for empowering Data Scientists and ML Engineers to unlock the full potential of their data. Currently serving as a valuable member of Voxel51, he takes a leading role in efforts to bridge the gap between practitioners and the necessary tools, enabling them to achieve exceptional outcomes. Daniel’s extensive experience in teaching and developing within the ML field has fueled his commitment to democratizing high-quality AI workflows for a wider audience.

April 24, 2025 - AI, Machine Learning and Computer Vision Meetup

April 24, 2025 - AI, Machine Learning and Computer Vision Meetup 2025-04-24 · 17:00

This is a virtual event.

Register for the Zoom

Towards a Multimodal AI Agent that Can See, Talk and Act

The development of multimodal AI agents marks a pivotal step toward creating systems capable of understanding, reasoning, and interacting with the world in human-like ways. Building such agents requires models that not only comprehend multi-sensory observations but also act adaptively to achieve goals within their environments. In this talk, I will present my research journey toward this grand goal across three key dimensions.

First, I will explore how to bridge the gap between core vision understanding and multimodal learning through unified frameworks at various granularities. Next, I will discuss connecting vision-language models with large language models (LLMs) to create intelligent conversational systems. Finally, I will delve into recent advancements that extend multimodal LLMs into vision-language-action models, forming the foundation for general-purpose robotics policies. To conclude, I will highlight ongoing efforts to develop agentic systems that integrate perception with action, enabling them to not only understand observations but also take meaningful actions in a single system.

Together, these lead to an aspiration of building the next generation of multimodal AI agents capable of seeing, talking, and acting across diverse scenarios in both digital and physical worlds.

About the Speaker

Jianwei Yang is a Principal Researcher at Microsoft Research (MSR), Redmond. His research focuses on the intersection of vision and multimodal learning, with an emphasis on bridging core vision tasks with language, building general-purpose and promptable multimodal models, and enabling these models to take meaningful actions in both virtual and physical environments.

ConceptAttention: Interpreting the Representations of Diffusion Transformers

Recently, diffusion transformers have taken over as the state-of-the-art model class for both image and video generation. However, similar to many existing deep learning architectures, their high-dimensional hidden representations are difficult to understand and interpret. This lack of interpretability is a barrier to their controllability and safe deployment.

We introduce ConceptAttention, an approach to interpreting the representations of diffusion transformers. Our method allows users to create rich saliency maps depicting the location and intensity of textual concepts. Our approach exposes how a diffusion model “sees” a generated image and notably requires no additional training. ConceptAttention improves upon widely used approaches like cross attention maps for isolating the location of visual concepts and even generalizes to real world (not just generated) images and video generation models!

Our work serves to improve the community’s understanding of how diffusion models represent data and has numerous potential applications, like image editing.

About the Speaker

Alec Helbling is a PhD student at Georgia Tech. His research focuses on improving the interpretability and controllability of generative models, particularly for image generation. His research is more application focused, and he has have interned at a variety of industrial research labs like Adobe Firefly, IBM Research, and NASA Jet Propulsion Lab. He also has a passion for creating explanatory videos of interesting machine learning and mathematical concepts.

RelationField: Relate Anything in Radiance Fields

Neural radiance fields recently emerged as a 3D scene representation extended by distilling open-vocabulary features from vision-language models. Current methods focus on object-centric tasks, leaving semantic relationships largely unexplored. We propose RelationField, the first method extracting inter-object relationships directly from neural radiance fields using pairs of rays for implicit relationship queries. RelationField distills relationship knowledge from multi-modal LLMs. Evaluated on open-vocabulary 3D scene graph generation and relationship-guided instance segmentation, RelationField achieves state-of-the-art performance.

About the Speaker

Sebastian Koch is a PhD student at Ulm University and Bosch Center for Artificial Intelligence. He is supervised by Timo Ropinski from Ulm University. His main research interest lies at the intersection of computer vision and robotics. The goal of his PhD is to develop 3D scene representations of the real world that are valuable for robots to navigate and solve tasks within their environment

RGB-X Model Development: Exploring Four Channel ML Workflows

Machine Learning is rapidly becoming multimodal. With many models in Computer Vision expanding to areas like vision and 3D, one area that has also quietly been advancing rapidly is RGB-X data, such as infrared, depth, or normals. In this talk we will cover some of the leading models in this exploding field of Visual AI and show some best practices on how to work with these complex data formats!

About the Speaker

Daniel Gural is a seasoned Machine Learning Evangelist with a strong passion for empowering Data Scientists and ML Engineers to unlock the full potential of their data. Currently serving as a valuable member of Voxel51, he takes a leading role in efforts to bridge the gap between practitioners and the necessary tools, enabling them to achieve exceptional outcomes. Daniel’s extensive experience in teaching and developing within the ML field has fueled his commitment to democratizing high-quality AI workflows for a wider audience.

April 24, 2025 - AI, Machine Learning and Computer Vision Meetup

April 24, 2025 - AI, Machine Learning and Computer Vision Meetup 2025-04-24 · 17:00

This is a virtual event.

Register for the Zoom

Towards a Multimodal AI Agent that Can See, Talk and Act

The development of multimodal AI agents marks a pivotal step toward creating systems capable of understanding, reasoning, and interacting with the world in human-like ways. Building such agents requires models that not only comprehend multi-sensory observations but also act adaptively to achieve goals within their environments. In this talk, I will present my research journey toward this grand goal across three key dimensions.

First, I will explore how to bridge the gap between core vision understanding and multimodal learning through unified frameworks at various granularities. Next, I will discuss connecting vision-language models with large language models (LLMs) to create intelligent conversational systems. Finally, I will delve into recent advancements that extend multimodal LLMs into vision-language-action models, forming the foundation for general-purpose robotics policies. To conclude, I will highlight ongoing efforts to develop agentic systems that integrate perception with action, enabling them to not only understand observations but also take meaningful actions in a single system.

Together, these lead to an aspiration of building the next generation of multimodal AI agents capable of seeing, talking, and acting across diverse scenarios in both digital and physical worlds.

About the Speaker

Jianwei Yang is a Principal Researcher at Microsoft Research (MSR), Redmond. His research focuses on the intersection of vision and multimodal learning, with an emphasis on bridging core vision tasks with language, building general-purpose and promptable multimodal models, and enabling these models to take meaningful actions in both virtual and physical environments.

ConceptAttention: Interpreting the Representations of Diffusion Transformers

Recently, diffusion transformers have taken over as the state-of-the-art model class for both image and video generation. However, similar to many existing deep learning architectures, their high-dimensional hidden representations are difficult to understand and interpret. This lack of interpretability is a barrier to their controllability and safe deployment.

We introduce ConceptAttention, an approach to interpreting the representations of diffusion transformers. Our method allows users to create rich saliency maps depicting the location and intensity of textual concepts. Our approach exposes how a diffusion model “sees” a generated image and notably requires no additional training. ConceptAttention improves upon widely used approaches like cross attention maps for isolating the location of visual concepts and even generalizes to real world (not just generated) images and video generation models!

Our work serves to improve the community’s understanding of how diffusion models represent data and has numerous potential applications, like image editing.

About the Speaker

Alec Helbling is a PhD student at Georgia Tech. His research focuses on improving the interpretability and controllability of generative models, particularly for image generation. His research is more application focused, and he has have interned at a variety of industrial research labs like Adobe Firefly, IBM Research, and NASA Jet Propulsion Lab. He also has a passion for creating explanatory videos of interesting machine learning and mathematical concepts.

RelationField: Relate Anything in Radiance Fields

Neural radiance fields recently emerged as a 3D scene representation extended by distilling open-vocabulary features from vision-language models. Current methods focus on object-centric tasks, leaving semantic relationships largely unexplored. We propose RelationField, the first method extracting inter-object relationships directly from neural radiance fields using pairs of rays for implicit relationship queries. RelationField distills relationship knowledge from multi-modal LLMs. Evaluated on open-vocabulary 3D scene graph generation and relationship-guided instance segmentation, RelationField achieves state-of-the-art performance.

About the Speaker

Sebastian Koch is a PhD student at Ulm University and Bosch Center for Artificial Intelligence. He is supervised by Timo Ropinski from Ulm University. His main research interest lies at the intersection of computer vision and robotics. The goal of his PhD is to develop 3D scene representations of the real world that are valuable for robots to navigate and solve tasks within their environment.

RGB-X Model Development: Exploring Four Channel ML Workflows

Machine Learning is rapidly becoming multimodal. With many models in Computer Vision expanding to areas like vision and 3D, one area that has also quietly been advancing rapidly is RGB-X data, such as infrared, depth, or normals. In this talk we will cover some of the leading models in this exploding field of Visual AI and show some best practices on how to work with these complex data formats!

About the Speaker

Daniel Gural is a seasoned Machine Learning Evangelist with a strong passion for empowering Data Scientists and ML Engineers to unlock the full potential of their data. Currently serving as a valuable member of Voxel51, he takes a leading role in efforts to bridge the gap between practitioners and the necessary tools, enabling them to achieve exceptional outcomes. Daniel’s extensive experience in teaching and developing within the ML field has fueled his commitment to democratizing high-quality AI workflows for a wider audience.

April 24, 2025 - AI, Machine Learning and Computer Vision Meetup

Generative AI Meetup (Paris) 2024-04-30 · 16:00

** Important RSVP HERE (Due to room capacity and venue security, it is required to pre-register at the link for admission)

Description: Welcome to the AI meetup in Paris. In collaboration with Roosh Circle, a European tech community for entrepreneurs, engineers, VC representatives, and business experts.

This meetup is devoted to Generative AI — the technology that is evolving rapidly and has great potential across diverse sectors. Let’s meet offline to discuss the power of GenAI across industries and strategies for integrating it into business.

Agenda: - 6:00-6:30pm: Gathering & Welcome Drinks - 6:30-6:40pm: Intro Word from Roosh Circle - 6:40-8:10pm: Tech talks and Q&A - 8:10-9:30pm: Networking & Drinks .

Tech Talk: Agile in the AI Era Speaker: Yassine AQEJJAJ (Devoteam) Abstract: In this talk, Yassine will dive deeper into the transformative impact of AI on Agile methodologies, focusing on software development.

Tech Talk: Scaling RAG: From Concept to Production Speaker: Louis GERMAIN (Ipsos Synthesio) Abstract: In this talk, Louise will share his experience with RAG evaluation and the challenges of scaling the feature with cost efficiency, stability, and speed.

Tech Talk: Advancing the LLM Deployment with Secure European Cloud Speaker: Frédéric BARDOLLE (Scaleway) Abstract: In this talk, Frédéric will share tactics for initiating secure deployment of AI projects, and shed light on how a resilient and sustainable cloud ecosystem can enhance LLMs utilization.

For whom: – AI/DL/ML Engineers with 3+ years of experience – ML researchers and professionals – Individuals with a solid foundation in GenAI – Product managers in AI-focused companies

When: April 30, 6 – 9 pm. Registration deadline: April 29

Venue: Liberte Living-Lab, 9 Rue d'Alexandrie, Atrium Space

Join Roosh Circle Community on Discord and LinkedIn

Event hosts: Roosh Circle and AICamp.

Let's meet and improve the world with the power of GenAI

Generative AI Meetup (Paris)

AI Meetup (April): Generative AI, LLMs and ML 2024-04-09 · 21:30

** Important RSVP here (Due to room capacity and building security, you must pre-register at the link for admission)

Description: Welcome to our in-person AI meetup in New York. Join us for deep dive tech talks on AI, GenAI, LLMs and ML, hands-on workshops, food/drink, networking with speakers and fellow developers.

Tech Talk: Build Lightweight LLMs Applications with Open Source Tools Speaker: Amogh Kokari (Con Edison) Abstract: ChatGPT is awesome, but developing with its API comes at a cost. Fortunately, there are open-source alternatives like Google Gemini, Streamlit, and Python APIs that can fetch prompt results using an API key. In this presentation, I'll explore how to create a lightweight, self-service end-to-end LLMs application using prompt engineering and fine-tuning based on user requests. Additionally, I'll demonstrate how to build a food suggestion application based on ingredients or food names.

Tech Talk: MultiModal Learning and Synthetic AI Speaker: Shafik Quoraishee (The New York Times) Abstract: This talk will be a Deep Dive into Multi-Modal AI Models, the powerful AI systems that are behind the functioning of applications such GPT Vision, DALL-E, and even Sora. We will go through the core theory of how intelligence from different sources of data (text, AI, and Vision) are combined to together in order help build capabilities like real-time Image/Video analysis, image hyper segmentation, and image to text extraction, and more.

Speakers/Topics: Stay tuned as we are updating speakers and schedules. If you have a keen interest in speaking to our community, we invite you to submit topics for consideration: Submit Topics

Sponsors: We are actively seeking sponsors to support our community. Whether it is by offering venue spaces, providing food/drink, or cash sponsor. Sponsors will not only speak at the meetups, receive prominent recognition, but also gain exposure to our extensive membership base of 20,000+ AI developers in New York or 350K+ worldwide.

Community on Slack/Discord

Event chat: chat and connect with speakers and attendees
Sharing blogs, events, job openings, projects collaborations
Join Slack/Discord (scroll down to the bottom) *

AI Meetup (April): Generative AI, LLMs and ML

talk-data.com

People (9 results)

Companies (1 result)

Activities & events