talk-data.com
People (19 results)
See all 19 →
Activities & events
| Title & Speakers | Event |
|---|---|
|
Nov 20 - Best of ICCV (Day 2)
2025-11-20 · 17:00
Welcome to the Best of ICCV series, your virtual pass to some of the groundbreaking research, insights, and innovations that defined this year’s conference. Live streaming from the authors to you. Date, Time and Location Nov 20, 2025 9 AM Pacific Online. Register for the Zoom! SGBD: Sharpness-Aware Mirror Gradient with BLIP-Based Denoising for Robust Multimodal Product Recommendation The growing integration of computer vision and machine learning into the retail industry—both online and in physical stores—has driven the adoption of multimodal recommender systems to help users navigate increasingly complex product landscapes. These systems leverage diverse data sources, such as product images, textual descriptions, and user-generated content, to better model user preferences and item characteristics. While the fusion of multimodal data helps address issues like data sparsity and cold-start problems, it also introduces challenges such as information inconsistency, noise, and increased training instability. In this paper, we analyze these robustness issues through the lens of flat local minima and propose a strategy that incorporates BLIP—a Vision-Language Model with strong denoising capabilities—to mitigate noise in multimodal inputs. Our method, Sharpness-Aware Mirror Gradient with BLIP-Based Denoising (SGBD), is a concise yet effective training strategy that implicitly enhances robustness during optimization. Extensive theoretical and empirical evaluations demonstrate its effectiveness across various multimodal recommendation benchmarks. SGBD offers a scalable solution for improving recommendation performance in real-world retail environments, where noisy, high-dimensional, and fast-evolving product data is the norm, making it a promising paradigm for training robust multi-modal recommender systems in retail industry. About the Speaker Kathy Wu holds a Ph.D. in Applied Mathematics and dual M.S. degrees in Computer Science and Quantitative Finance from the University of Southern California (USC), Los Angeles, CA, USA. At USC, she served as a course lecturer, offering ML Foundations and ML for Business Applications in the science school and business school. Her academic research spans high-dimensional statistics, deep learning, and causal inference, etc. Kathy brings industry experience from Meta, LinkedIn, and Morgan Stanley in the Bay Area and New York City, US, where she focused on AI methodologies and real-world applications. She is currently an Applied Scientist at Amazon, within the Global Store organization, leading projects in E-Commerce Recommendation Systems, Search Engines, Multi-Modal Vision-Language Models (VLMs), and LLM/GenAI in retails. Her work has been published in top-tier conferences including ICCV, CVPR, ICLR, SIGIR, WACV, etc. At ICCV 2025, she won the Best Paper Award in Retail Vision. Spatial Mental Modeling from Limited Views Can VLMs imagine the unobservable space from just a few views, like humans do? Humans form spatial mental models, as internal representations of "unseen space" to reason about layout, perspective, and motion. On our proposed MINDCUBE, we see critical gap systematically on VLMs building robust spatial mental models through representing positions (cognitive mapping), orientations (perspective-taking), and dynamics (mental simulation for ''what-if'' movements). We then explore three approaches to help VLMs approximate spatial mental models, including unseen intermediate views, natural language reasoning chains, and cognitive maps. The significant improvement comes from ''map-then-reason'' that jointly trains the model to first abstract a cognitive map and then reason upon it. By training models to construct and reason over these internal maps, we boosted accuracy from 37.8% to 60.8% (+23.0%). Adding reinforcement learning pushed performance even further to 70.7% (+32.9%). Our key insight is that such scaffolding of spatial mental models, actively constructing and utilizing internal structured spatial representations with flexible reasoning processes, significantly improves understanding of "unobservable space". We aim to understand why geometric concepts remain challenging for VLMs and outlining promising research directions towards fostering more robust spatial intelligence. About the Speaker Manling Li is an Assistant Professor at Northwestern University and Amazon Scholar. She was a postdoc at Stanford University, and obtained the PhD degree in Computer Science at University of Illinois Urbana-Champaign in 2023. She works on the intersection of language, vision, and robotics, recognized by the MIT TR 35 Under 35, ACL Inaugural Dissertation Award Honorable Mention, ACL’24 Outstanding Paper Award, ACL'20 Best Demo Paper Award, and NAACL'21 Best Demo Paper Award, Microsoft Research PhD Fellowship, EE CS Rising Star, etc. Forecasting and Visualizing Air Pollution via Sky Images and VLM-Guided Generative Models Air pollution monitoring is traditionally limited by costly sensors and sparse data coverage. Our research introduces a vision-language model framework that predicts air quality directly from real-world sky images and also simulates skies under varying pollution levels to enhance interpretability and robustness. We further develop visualization techniques to make predictions more understandable for policymakers and the public. This talk will present our methodology, key findings, and implications for sustainable urban environments. About the Speaker Mohammad Saleh Vahdatpour is a PhD candidate in Computer Science at Georgia State University specializing in deep learning, vision–language models, and sustainable AI systems. His research bridges generative AI, environmental monitoring, and motion perception, focusing on scalable and energy-efficient models that connect scientific innovation with real-world impact. Sari Sandbox: A Virtual Retail Store Environment for Embodied AI Agents We present Sari Sandbox, a high-fidelity, photorealistic 3D retail store simulation for benchmarking embodied agents against human performance in shopping tasks. Addressing a gap in retail-specific sim environments for embodied agent training, Sari Sandbox features over 250 interactive grocery items across three store configurations, controlled via an API. It supports both virtual reality (VR) for human interaction and a vision language model (VLM)-powered embodied agent. We also introduce SariBench, a dataset of annotated human demonstrations across varied task difficulties. Our sandbox enables embodied agents to navigate, inspect, and manipulate retail items, providing baselines against human performance. We conclude with benchmarks, performance analysis, and recommendations for enhancing realism and scalability. About the Speakers Emmanuel G. Maminta is a fourth-year Artificial Intelligence Ph.D. student at the Ubiquitous Computing Laboratory (UCL) in the University of the Philippines Diliman, advised by Prof. Rowel O. Atienza. Janika Deborah B.Gajo is an undergraduate student studying for a Bachelor of Science in Computer Engineering at the University of the Philippines, Diliman. |
Nov 20 - Best of ICCV (Day 2)
|
|
Nov 20 - Best of ICCV (Day 2)
2025-11-20 · 17:00
Welcome to the Best of ICCV series, your virtual pass to some of the groundbreaking research, insights, and innovations that defined this year’s conference. Live streaming from the authors to you. Date, Time and Location Nov 20, 2025 9 AM Pacific Online. Register for the Zoom! SGBD: Sharpness-Aware Mirror Gradient with BLIP-Based Denoising for Robust Multimodal Product Recommendation The growing integration of computer vision and machine learning into the retail industry—both online and in physical stores—has driven the adoption of multimodal recommender systems to help users navigate increasingly complex product landscapes. These systems leverage diverse data sources, such as product images, textual descriptions, and user-generated content, to better model user preferences and item characteristics. While the fusion of multimodal data helps address issues like data sparsity and cold-start problems, it also introduces challenges such as information inconsistency, noise, and increased training instability. In this paper, we analyze these robustness issues through the lens of flat local minima and propose a strategy that incorporates BLIP—a Vision-Language Model with strong denoising capabilities—to mitigate noise in multimodal inputs. Our method, Sharpness-Aware Mirror Gradient with BLIP-Based Denoising (SGBD), is a concise yet effective training strategy that implicitly enhances robustness during optimization. Extensive theoretical and empirical evaluations demonstrate its effectiveness across various multimodal recommendation benchmarks. SGBD offers a scalable solution for improving recommendation performance in real-world retail environments, where noisy, high-dimensional, and fast-evolving product data is the norm, making it a promising paradigm for training robust multi-modal recommender systems in retail industry. About the Speaker Kathy Wu holds a Ph.D. in Applied Mathematics and dual M.S. degrees in Computer Science and Quantitative Finance from the University of Southern California (USC), Los Angeles, CA, USA. At USC, she served as a course lecturer, offering ML Foundations and ML for Business Applications in the science school and business school. Her academic research spans high-dimensional statistics, deep learning, and causal inference, etc. Kathy brings industry experience from Meta, LinkedIn, and Morgan Stanley in the Bay Area and New York City, US, where she focused on AI methodologies and real-world applications. She is currently an Applied Scientist at Amazon, within the Global Store organization, leading projects in E-Commerce Recommendation Systems, Search Engines, Multi-Modal Vision-Language Models (VLMs), and LLM/GenAI in retails. Her work has been published in top-tier conferences including ICCV, CVPR, ICLR, SIGIR, WACV, etc. At ICCV 2025, she won the Best Paper Award in Retail Vision. Spatial Mental Modeling from Limited Views Can VLMs imagine the unobservable space from just a few views, like humans do? Humans form spatial mental models, as internal representations of "unseen space" to reason about layout, perspective, and motion. On our proposed MINDCUBE, we see critical gap systematically on VLMs building robust spatial mental models through representing positions (cognitive mapping), orientations (perspective-taking), and dynamics (mental simulation for ''what-if'' movements). We then explore three approaches to help VLMs approximate spatial mental models, including unseen intermediate views, natural language reasoning chains, and cognitive maps. The significant improvement comes from ''map-then-reason'' that jointly trains the model to first abstract a cognitive map and then reason upon it. By training models to construct and reason over these internal maps, we boosted accuracy from 37.8% to 60.8% (+23.0%). Adding reinforcement learning pushed performance even further to 70.7% (+32.9%). Our key insight is that such scaffolding of spatial mental models, actively constructing and utilizing internal structured spatial representations with flexible reasoning processes, significantly improves understanding of "unobservable space". We aim to understand why geometric concepts remain challenging for VLMs and outlining promising research directions towards fostering more robust spatial intelligence. About the Speaker Manling Li is an Assistant Professor at Northwestern University and Amazon Scholar. She was a postdoc at Stanford University, and obtained the PhD degree in Computer Science at University of Illinois Urbana-Champaign in 2023. She works on the intersection of language, vision, and robotics, recognized by the MIT TR 35 Under 35, ACL Inaugural Dissertation Award Honorable Mention, ACL’24 Outstanding Paper Award, ACL'20 Best Demo Paper Award, and NAACL'21 Best Demo Paper Award, Microsoft Research PhD Fellowship, EE CS Rising Star, etc. Forecasting and Visualizing Air Pollution via Sky Images and VLM-Guided Generative Models Air pollution monitoring is traditionally limited by costly sensors and sparse data coverage. Our research introduces a vision-language model framework that predicts air quality directly from real-world sky images and also simulates skies under varying pollution levels to enhance interpretability and robustness. We further develop visualization techniques to make predictions more understandable for policymakers and the public. This talk will present our methodology, key findings, and implications for sustainable urban environments. About the Speaker Mohammad Saleh Vahdatpour is a PhD candidate in Computer Science at Georgia State University specializing in deep learning, vision–language models, and sustainable AI systems. His research bridges generative AI, environmental monitoring, and motion perception, focusing on scalable and energy-efficient models that connect scientific innovation with real-world impact. Sari Sandbox: A Virtual Retail Store Environment for Embodied AI Agents We present Sari Sandbox, a high-fidelity, photorealistic 3D retail store simulation for benchmarking embodied agents against human performance in shopping tasks. Addressing a gap in retail-specific sim environments for embodied agent training, Sari Sandbox features over 250 interactive grocery items across three store configurations, controlled via an API. It supports both virtual reality (VR) for human interaction and a vision language model (VLM)-powered embodied agent. We also introduce SariBench, a dataset of annotated human demonstrations across varied task difficulties. Our sandbox enables embodied agents to navigate, inspect, and manipulate retail items, providing baselines against human performance. We conclude with benchmarks, performance analysis, and recommendations for enhancing realism and scalability. About the Speakers Emmanuel G. Maminta is a fourth-year Artificial Intelligence Ph.D. student at the Ubiquitous Computing Laboratory (UCL) in the University of the Philippines Diliman, advised by Prof. Rowel O. Atienza. Janika Deborah B.Gajo is an undergraduate student studying for a Bachelor of Science in Computer Engineering at the University of the Philippines, Diliman. |
Nov 20 - Best of ICCV (Day 2)
|
|
Nov 20 - Best of ICCV (Day 2)
2025-11-20 · 17:00
Welcome to the Best of ICCV series, your virtual pass to some of the groundbreaking research, insights, and innovations that defined this year’s conference. Live streaming from the authors to you. Date, Time and Location Nov 20, 2025 9 AM Pacific Online. Register for the Zoom! SGBD: Sharpness-Aware Mirror Gradient with BLIP-Based Denoising for Robust Multimodal Product Recommendation The growing integration of computer vision and machine learning into the retail industry—both online and in physical stores—has driven the adoption of multimodal recommender systems to help users navigate increasingly complex product landscapes. These systems leverage diverse data sources, such as product images, textual descriptions, and user-generated content, to better model user preferences and item characteristics. While the fusion of multimodal data helps address issues like data sparsity and cold-start problems, it also introduces challenges such as information inconsistency, noise, and increased training instability. In this paper, we analyze these robustness issues through the lens of flat local minima and propose a strategy that incorporates BLIP—a Vision-Language Model with strong denoising capabilities—to mitigate noise in multimodal inputs. Our method, Sharpness-Aware Mirror Gradient with BLIP-Based Denoising (SGBD), is a concise yet effective training strategy that implicitly enhances robustness during optimization. Extensive theoretical and empirical evaluations demonstrate its effectiveness across various multimodal recommendation benchmarks. SGBD offers a scalable solution for improving recommendation performance in real-world retail environments, where noisy, high-dimensional, and fast-evolving product data is the norm, making it a promising paradigm for training robust multi-modal recommender systems in retail industry. About the Speaker Kathy Wu holds a Ph.D. in Applied Mathematics and dual M.S. degrees in Computer Science and Quantitative Finance from the University of Southern California (USC), Los Angeles, CA, USA. At USC, she served as a course lecturer, offering ML Foundations and ML for Business Applications in the science school and business school. Her academic research spans high-dimensional statistics, deep learning, and causal inference, etc. Kathy brings industry experience from Meta, LinkedIn, and Morgan Stanley in the Bay Area and New York City, US, where she focused on AI methodologies and real-world applications. She is currently an Applied Scientist at Amazon, within the Global Store organization, leading projects in E-Commerce Recommendation Systems, Search Engines, Multi-Modal Vision-Language Models (VLMs), and LLM/GenAI in retails. Her work has been published in top-tier conferences including ICCV, CVPR, ICLR, SIGIR, WACV, etc. At ICCV 2025, she won the Best Paper Award in Retail Vision. Spatial Mental Modeling from Limited Views Can VLMs imagine the unobservable space from just a few views, like humans do? Humans form spatial mental models, as internal representations of "unseen space" to reason about layout, perspective, and motion. On our proposed MINDCUBE, we see critical gap systematically on VLMs building robust spatial mental models through representing positions (cognitive mapping), orientations (perspective-taking), and dynamics (mental simulation for ''what-if'' movements). We then explore three approaches to help VLMs approximate spatial mental models, including unseen intermediate views, natural language reasoning chains, and cognitive maps. The significant improvement comes from ''map-then-reason'' that jointly trains the model to first abstract a cognitive map and then reason upon it. By training models to construct and reason over these internal maps, we boosted accuracy from 37.8% to 60.8% (+23.0%). Adding reinforcement learning pushed performance even further to 70.7% (+32.9%). Our key insight is that such scaffolding of spatial mental models, actively constructing and utilizing internal structured spatial representations with flexible reasoning processes, significantly improves understanding of "unobservable space". We aim to understand why geometric concepts remain challenging for VLMs and outlining promising research directions towards fostering more robust spatial intelligence. About the Speaker Manling Li is an Assistant Professor at Northwestern University and Amazon Scholar. She was a postdoc at Stanford University, and obtained the PhD degree in Computer Science at University of Illinois Urbana-Champaign in 2023. She works on the intersection of language, vision, and robotics, recognized by the MIT TR 35 Under 35, ACL Inaugural Dissertation Award Honorable Mention, ACL’24 Outstanding Paper Award, ACL'20 Best Demo Paper Award, and NAACL'21 Best Demo Paper Award, Microsoft Research PhD Fellowship, EE CS Rising Star, etc. Forecasting and Visualizing Air Pollution via Sky Images and VLM-Guided Generative Models Air pollution monitoring is traditionally limited by costly sensors and sparse data coverage. Our research introduces a vision-language model framework that predicts air quality directly from real-world sky images and also simulates skies under varying pollution levels to enhance interpretability and robustness. We further develop visualization techniques to make predictions more understandable for policymakers and the public. This talk will present our methodology, key findings, and implications for sustainable urban environments. About the Speaker Mohammad Saleh Vahdatpour is a PhD candidate in Computer Science at Georgia State University specializing in deep learning, vision–language models, and sustainable AI systems. His research bridges generative AI, environmental monitoring, and motion perception, focusing on scalable and energy-efficient models that connect scientific innovation with real-world impact. Sari Sandbox: A Virtual Retail Store Environment for Embodied AI Agents We present Sari Sandbox, a high-fidelity, photorealistic 3D retail store simulation for benchmarking embodied agents against human performance in shopping tasks. Addressing a gap in retail-specific sim environments for embodied agent training, Sari Sandbox features over 250 interactive grocery items across three store configurations, controlled via an API. It supports both virtual reality (VR) for human interaction and a vision language model (VLM)-powered embodied agent. We also introduce SariBench, a dataset of annotated human demonstrations across varied task difficulties. Our sandbox enables embodied agents to navigate, inspect, and manipulate retail items, providing baselines against human performance. We conclude with benchmarks, performance analysis, and recommendations for enhancing realism and scalability. About the Speakers Emmanuel G. Maminta is a fourth-year Artificial Intelligence Ph.D. student at the Ubiquitous Computing Laboratory (UCL) in the University of the Philippines Diliman, advised by Prof. Rowel O. Atienza. Janika Deborah B.Gajo is an undergraduate student studying for a Bachelor of Science in Computer Engineering at the University of the Philippines, Diliman. |
Nov 20 - Best of ICCV (Day 2)
|
|
Nov 20 - Best of ICCV (Day 2)
2025-11-20 · 17:00
Welcome to the Best of ICCV series, your virtual pass to some of the groundbreaking research, insights, and innovations that defined this year’s conference. Live streaming from the authors to you. Date, Time and Location Nov 20, 2025 9 AM Pacific Online. Register for the Zoom! SGBD: Sharpness-Aware Mirror Gradient with BLIP-Based Denoising for Robust Multimodal Product Recommendation The growing integration of computer vision and machine learning into the retail industry—both online and in physical stores—has driven the adoption of multimodal recommender systems to help users navigate increasingly complex product landscapes. These systems leverage diverse data sources, such as product images, textual descriptions, and user-generated content, to better model user preferences and item characteristics. While the fusion of multimodal data helps address issues like data sparsity and cold-start problems, it also introduces challenges such as information inconsistency, noise, and increased training instability. In this paper, we analyze these robustness issues through the lens of flat local minima and propose a strategy that incorporates BLIP—a Vision-Language Model with strong denoising capabilities—to mitigate noise in multimodal inputs. Our method, Sharpness-Aware Mirror Gradient with BLIP-Based Denoising (SGBD), is a concise yet effective training strategy that implicitly enhances robustness during optimization. Extensive theoretical and empirical evaluations demonstrate its effectiveness across various multimodal recommendation benchmarks. SGBD offers a scalable solution for improving recommendation performance in real-world retail environments, where noisy, high-dimensional, and fast-evolving product data is the norm, making it a promising paradigm for training robust multi-modal recommender systems in retail industry. About the Speaker Kathy Wu holds a Ph.D. in Applied Mathematics and dual M.S. degrees in Computer Science and Quantitative Finance from the University of Southern California (USC), Los Angeles, CA, USA. At USC, she served as a course lecturer, offering ML Foundations and ML for Business Applications in the science school and business school. Her academic research spans high-dimensional statistics, deep learning, and causal inference, etc. Kathy brings industry experience from Meta, LinkedIn, and Morgan Stanley in the Bay Area and New York City, US, where she focused on AI methodologies and real-world applications. She is currently an Applied Scientist at Amazon, within the Global Store organization, leading projects in E-Commerce Recommendation Systems, Search Engines, Multi-Modal Vision-Language Models (VLMs), and LLM/GenAI in retails. Her work has been published in top-tier conferences including ICCV, CVPR, ICLR, SIGIR, WACV, etc. At ICCV 2025, she won the Best Paper Award in Retail Vision. Spatial Mental Modeling from Limited Views Can VLMs imagine the unobservable space from just a few views, like humans do? Humans form spatial mental models, as internal representations of "unseen space" to reason about layout, perspective, and motion. On our proposed MINDCUBE, we see critical gap systematically on VLMs building robust spatial mental models through representing positions (cognitive mapping), orientations (perspective-taking), and dynamics (mental simulation for ''what-if'' movements). We then explore three approaches to help VLMs approximate spatial mental models, including unseen intermediate views, natural language reasoning chains, and cognitive maps. The significant improvement comes from ''map-then-reason'' that jointly trains the model to first abstract a cognitive map and then reason upon it. By training models to construct and reason over these internal maps, we boosted accuracy from 37.8% to 60.8% (+23.0%). Adding reinforcement learning pushed performance even further to 70.7% (+32.9%). Our key insight is that such scaffolding of spatial mental models, actively constructing and utilizing internal structured spatial representations with flexible reasoning processes, significantly improves understanding of "unobservable space". We aim to understand why geometric concepts remain challenging for VLMs and outlining promising research directions towards fostering more robust spatial intelligence. About the Speaker Manling Li is an Assistant Professor at Northwestern University and Amazon Scholar. She was a postdoc at Stanford University, and obtained the PhD degree in Computer Science at University of Illinois Urbana-Champaign in 2023. She works on the intersection of language, vision, and robotics, recognized by the MIT TR 35 Under 35, ACL Inaugural Dissertation Award Honorable Mention, ACL’24 Outstanding Paper Award, ACL'20 Best Demo Paper Award, and NAACL'21 Best Demo Paper Award, Microsoft Research PhD Fellowship, EE CS Rising Star, etc. Forecasting and Visualizing Air Pollution via Sky Images and VLM-Guided Generative Models Air pollution monitoring is traditionally limited by costly sensors and sparse data coverage. Our research introduces a vision-language model framework that predicts air quality directly from real-world sky images and also simulates skies under varying pollution levels to enhance interpretability and robustness. We further develop visualization techniques to make predictions more understandable for policymakers and the public. This talk will present our methodology, key findings, and implications for sustainable urban environments. About the Speaker Mohammad Saleh Vahdatpour is a PhD candidate in Computer Science at Georgia State University specializing in deep learning, vision–language models, and sustainable AI systems. His research bridges generative AI, environmental monitoring, and motion perception, focusing on scalable and energy-efficient models that connect scientific innovation with real-world impact. Sari Sandbox: A Virtual Retail Store Environment for Embodied AI Agents We present Sari Sandbox, a high-fidelity, photorealistic 3D retail store simulation for benchmarking embodied agents against human performance in shopping tasks. Addressing a gap in retail-specific sim environments for embodied agent training, Sari Sandbox features over 250 interactive grocery items across three store configurations, controlled via an API. It supports both virtual reality (VR) for human interaction and a vision language model (VLM)-powered embodied agent. We also introduce SariBench, a dataset of annotated human demonstrations across varied task difficulties. Our sandbox enables embodied agents to navigate, inspect, and manipulate retail items, providing baselines against human performance. We conclude with benchmarks, performance analysis, and recommendations for enhancing realism and scalability. About the Speakers Emmanuel G. Maminta is a fourth-year Artificial Intelligence Ph.D. student at the Ubiquitous Computing Laboratory (UCL) in the University of the Philippines Diliman, advised by Prof. Rowel O. Atienza. Janika Deborah B.Gajo is an undergraduate student studying for a Bachelor of Science in Computer Engineering at the University of the Philippines, Diliman. |
Nov 20 - Best of ICCV (Day 2)
|
|
Nov 20 - Best of ICCV (Day 2)
2025-11-20 · 17:00
Welcome to the Best of ICCV series, your virtual pass to some of the groundbreaking research, insights, and innovations that defined this year’s conference. Live streaming from the authors to you. Date, Time and Location Nov 20, 2025 9 AM Pacific Online. Register for the Zoom! SGBD: Sharpness-Aware Mirror Gradient with BLIP-Based Denoising for Robust Multimodal Product Recommendation The growing integration of computer vision and machine learning into the retail industry—both online and in physical stores—has driven the adoption of multimodal recommender systems to help users navigate increasingly complex product landscapes. These systems leverage diverse data sources, such as product images, textual descriptions, and user-generated content, to better model user preferences and item characteristics. While the fusion of multimodal data helps address issues like data sparsity and cold-start problems, it also introduces challenges such as information inconsistency, noise, and increased training instability. In this paper, we analyze these robustness issues through the lens of flat local minima and propose a strategy that incorporates BLIP—a Vision-Language Model with strong denoising capabilities—to mitigate noise in multimodal inputs. Our method, Sharpness-Aware Mirror Gradient with BLIP-Based Denoising (SGBD), is a concise yet effective training strategy that implicitly enhances robustness during optimization. Extensive theoretical and empirical evaluations demonstrate its effectiveness across various multimodal recommendation benchmarks. SGBD offers a scalable solution for improving recommendation performance in real-world retail environments, where noisy, high-dimensional, and fast-evolving product data is the norm, making it a promising paradigm for training robust multi-modal recommender systems in retail industry. About the Speaker Kathy Wu holds a Ph.D. in Applied Mathematics and dual M.S. degrees in Computer Science and Quantitative Finance from the University of Southern California (USC), Los Angeles, CA, USA. At USC, she served as a course lecturer, offering ML Foundations and ML for Business Applications in the science school and business school. Her academic research spans high-dimensional statistics, deep learning, and causal inference, etc. Kathy brings industry experience from Meta, LinkedIn, and Morgan Stanley in the Bay Area and New York City, US, where she focused on AI methodologies and real-world applications. She is currently an Applied Scientist at Amazon, within the Global Store organization, leading projects in E-Commerce Recommendation Systems, Search Engines, Multi-Modal Vision-Language Models (VLMs), and LLM/GenAI in retails. Her work has been published in top-tier conferences including ICCV, CVPR, ICLR, SIGIR, WACV, etc. At ICCV 2025, she won the Best Paper Award in Retail Vision. Spatial Mental Modeling from Limited Views Can VLMs imagine the unobservable space from just a few views, like humans do? Humans form spatial mental models, as internal representations of "unseen space" to reason about layout, perspective, and motion. On our proposed MINDCUBE, we see critical gap systematically on VLMs building robust spatial mental models through representing positions (cognitive mapping), orientations (perspective-taking), and dynamics (mental simulation for ''what-if'' movements). We then explore three approaches to help VLMs approximate spatial mental models, including unseen intermediate views, natural language reasoning chains, and cognitive maps. The significant improvement comes from ''map-then-reason'' that jointly trains the model to first abstract a cognitive map and then reason upon it. By training models to construct and reason over these internal maps, we boosted accuracy from 37.8% to 60.8% (+23.0%). Adding reinforcement learning pushed performance even further to 70.7% (+32.9%). Our key insight is that such scaffolding of spatial mental models, actively constructing and utilizing internal structured spatial representations with flexible reasoning processes, significantly improves understanding of "unobservable space". We aim to understand why geometric concepts remain challenging for VLMs and outlining promising research directions towards fostering more robust spatial intelligence. About the Speaker Manling Li is an Assistant Professor at Northwestern University and Amazon Scholar. She was a postdoc at Stanford University, and obtained the PhD degree in Computer Science at University of Illinois Urbana-Champaign in 2023. She works on the intersection of language, vision, and robotics, recognized by the MIT TR 35 Under 35, ACL Inaugural Dissertation Award Honorable Mention, ACL’24 Outstanding Paper Award, ACL'20 Best Demo Paper Award, and NAACL'21 Best Demo Paper Award, Microsoft Research PhD Fellowship, EE CS Rising Star, etc. Forecasting and Visualizing Air Pollution via Sky Images and VLM-Guided Generative Models Air pollution monitoring is traditionally limited by costly sensors and sparse data coverage. Our research introduces a vision-language model framework that predicts air quality directly from real-world sky images and also simulates skies under varying pollution levels to enhance interpretability and robustness. We further develop visualization techniques to make predictions more understandable for policymakers and the public. This talk will present our methodology, key findings, and implications for sustainable urban environments. About the Speaker Mohammad Saleh Vahdatpour is a PhD candidate in Computer Science at Georgia State University specializing in deep learning, vision–language models, and sustainable AI systems. His research bridges generative AI, environmental monitoring, and motion perception, focusing on scalable and energy-efficient models that connect scientific innovation with real-world impact. Sari Sandbox: A Virtual Retail Store Environment for Embodied AI Agents We present Sari Sandbox, a high-fidelity, photorealistic 3D retail store simulation for benchmarking embodied agents against human performance in shopping tasks. Addressing a gap in retail-specific sim environments for embodied agent training, Sari Sandbox features over 250 interactive grocery items across three store configurations, controlled via an API. It supports both virtual reality (VR) for human interaction and a vision language model (VLM)-powered embodied agent. We also introduce SariBench, a dataset of annotated human demonstrations across varied task difficulties. Our sandbox enables embodied agents to navigate, inspect, and manipulate retail items, providing baselines against human performance. We conclude with benchmarks, performance analysis, and recommendations for enhancing realism and scalability. About the Speakers Emmanuel G. Maminta is a fourth-year Artificial Intelligence Ph.D. student at the Ubiquitous Computing Laboratory (UCL) in the University of the Philippines Diliman, advised by Prof. Rowel O. Atienza. Janika Deborah B.Gajo is an undergraduate student studying for a Bachelor of Science in Computer Engineering at the University of the Philippines, Diliman. |
Nov 20 - Best of ICCV (Day 2)
|
|
Nov 20 - Best of ICCV (Day 2)
2025-11-20 · 17:00
Welcome to the Best of ICCV series, your virtual pass to some of the groundbreaking research, insights, and innovations that defined this year’s conference. Live streaming from the authors to you. Date, Time and Location Nov 20, 2025 9 AM Pacific Online. Register for the Zoom! SGBD: Sharpness-Aware Mirror Gradient with BLIP-Based Denoising for Robust Multimodal Product Recommendation The growing integration of computer vision and machine learning into the retail industry—both online and in physical stores—has driven the adoption of multimodal recommender systems to help users navigate increasingly complex product landscapes. These systems leverage diverse data sources, such as product images, textual descriptions, and user-generated content, to better model user preferences and item characteristics. While the fusion of multimodal data helps address issues like data sparsity and cold-start problems, it also introduces challenges such as information inconsistency, noise, and increased training instability. In this paper, we analyze these robustness issues through the lens of flat local minima and propose a strategy that incorporates BLIP—a Vision-Language Model with strong denoising capabilities—to mitigate noise in multimodal inputs. Our method, Sharpness-Aware Mirror Gradient with BLIP-Based Denoising (SGBD), is a concise yet effective training strategy that implicitly enhances robustness during optimization. Extensive theoretical and empirical evaluations demonstrate its effectiveness across various multimodal recommendation benchmarks. SGBD offers a scalable solution for improving recommendation performance in real-world retail environments, where noisy, high-dimensional, and fast-evolving product data is the norm, making it a promising paradigm for training robust multi-modal recommender systems in retail industry. About the Speaker Kathy Wu holds a Ph.D. in Applied Mathematics and dual M.S. degrees in Computer Science and Quantitative Finance from the University of Southern California (USC), Los Angeles, CA, USA. At USC, she served as a course lecturer, offering ML Foundations and ML for Business Applications in the science school and business school. Her academic research spans high-dimensional statistics, deep learning, and causal inference, etc. Kathy brings industry experience from Meta, LinkedIn, and Morgan Stanley in the Bay Area and New York City, US, where she focused on AI methodologies and real-world applications. She is currently an Applied Scientist at Amazon, within the Global Store organization, leading projects in E-Commerce Recommendation Systems, Search Engines, Multi-Modal Vision-Language Models (VLMs), and LLM/GenAI in retails. Her work has been published in top-tier conferences including ICCV, CVPR, ICLR, SIGIR, WACV, etc. At ICCV 2025, she won the Best Paper Award in Retail Vision. Spatial Mental Modeling from Limited Views Can VLMs imagine the unobservable space from just a few views, like humans do? Humans form spatial mental models, as internal representations of "unseen space" to reason about layout, perspective, and motion. On our proposed MINDCUBE, we see critical gap systematically on VLMs building robust spatial mental models through representing positions (cognitive mapping), orientations (perspective-taking), and dynamics (mental simulation for ''what-if'' movements). We then explore three approaches to help VLMs approximate spatial mental models, including unseen intermediate views, natural language reasoning chains, and cognitive maps. The significant improvement comes from ''map-then-reason'' that jointly trains the model to first abstract a cognitive map and then reason upon it. By training models to construct and reason over these internal maps, we boosted accuracy from 37.8% to 60.8% (+23.0%). Adding reinforcement learning pushed performance even further to 70.7% (+32.9%). Our key insight is that such scaffolding of spatial mental models, actively constructing and utilizing internal structured spatial representations with flexible reasoning processes, significantly improves understanding of "unobservable space". We aim to understand why geometric concepts remain challenging for VLMs and outlining promising research directions towards fostering more robust spatial intelligence. About the Speaker Manling Li is an Assistant Professor at Northwestern University and Amazon Scholar. She was a postdoc at Stanford University, and obtained the PhD degree in Computer Science at University of Illinois Urbana-Champaign in 2023. She works on the intersection of language, vision, and robotics, recognized by the MIT TR 35 Under 35, ACL Inaugural Dissertation Award Honorable Mention, ACL’24 Outstanding Paper Award, ACL'20 Best Demo Paper Award, and NAACL'21 Best Demo Paper Award, Microsoft Research PhD Fellowship, EE CS Rising Star, etc. Forecasting and Visualizing Air Pollution via Sky Images and VLM-Guided Generative Models Air pollution monitoring is traditionally limited by costly sensors and sparse data coverage. Our research introduces a vision-language model framework that predicts air quality directly from real-world sky images and also simulates skies under varying pollution levels to enhance interpretability and robustness. We further develop visualization techniques to make predictions more understandable for policymakers and the public. This talk will present our methodology, key findings, and implications for sustainable urban environments. About the Speaker Mohammad Saleh Vahdatpour is a PhD candidate in Computer Science at Georgia State University specializing in deep learning, vision–language models, and sustainable AI systems. His research bridges generative AI, environmental monitoring, and motion perception, focusing on scalable and energy-efficient models that connect scientific innovation with real-world impact. Sari Sandbox: A Virtual Retail Store Environment for Embodied AI Agents We present Sari Sandbox, a high-fidelity, photorealistic 3D retail store simulation for benchmarking embodied agents against human performance in shopping tasks. Addressing a gap in retail-specific sim environments for embodied agent training, Sari Sandbox features over 250 interactive grocery items across three store configurations, controlled via an API. It supports both virtual reality (VR) for human interaction and a vision language model (VLM)-powered embodied agent. We also introduce SariBench, a dataset of annotated human demonstrations across varied task difficulties. Our sandbox enables embodied agents to navigate, inspect, and manipulate retail items, providing baselines against human performance. We conclude with benchmarks, performance analysis, and recommendations for enhancing realism and scalability. About the Speakers Emmanuel G. Maminta is a fourth-year Artificial Intelligence Ph.D. student at the Ubiquitous Computing Laboratory (UCL) in the University of the Philippines Diliman, advised by Prof. Rowel O. Atienza. Janika Deborah B.Gajo is an undergraduate student studying for a Bachelor of Science in Computer Engineering at the University of the Philippines, Diliman. |
Nov 20 - Best of ICCV (Day 2)
|
|
Nov 20 - Best of ICCV (Day 2)
2025-11-20 · 17:00
Welcome to the Best of ICCV series, your virtual pass to some of the groundbreaking research, insights, and innovations that defined this year’s conference. Live streaming from the authors to you. Date, Time and Location Nov 20, 2025 9 AM Pacific Online. Register for the Zoom! SGBD: Sharpness-Aware Mirror Gradient with BLIP-Based Denoising for Robust Multimodal Product Recommendation The growing integration of computer vision and machine learning into the retail industry—both online and in physical stores—has driven the adoption of multimodal recommender systems to help users navigate increasingly complex product landscapes. These systems leverage diverse data sources, such as product images, textual descriptions, and user-generated content, to better model user preferences and item characteristics. While the fusion of multimodal data helps address issues like data sparsity and cold-start problems, it also introduces challenges such as information inconsistency, noise, and increased training instability. In this paper, we analyze these robustness issues through the lens of flat local minima and propose a strategy that incorporates BLIP—a Vision-Language Model with strong denoising capabilities—to mitigate noise in multimodal inputs. Our method, Sharpness-Aware Mirror Gradient with BLIP-Based Denoising (SGBD), is a concise yet effective training strategy that implicitly enhances robustness during optimization. Extensive theoretical and empirical evaluations demonstrate its effectiveness across various multimodal recommendation benchmarks. SGBD offers a scalable solution for improving recommendation performance in real-world retail environments, where noisy, high-dimensional, and fast-evolving product data is the norm, making it a promising paradigm for training robust multi-modal recommender systems in retail industry. About the Speaker Kathy Wu holds a Ph.D. in Applied Mathematics and dual M.S. degrees in Computer Science and Quantitative Finance from the University of Southern California (USC), Los Angeles, CA, USA. At USC, she served as a course lecturer, offering ML Foundations and ML for Business Applications in the science school and business school. Her academic research spans high-dimensional statistics, deep learning, and causal inference, etc. Kathy brings industry experience from Meta, LinkedIn, and Morgan Stanley in the Bay Area and New York City, US, where she focused on AI methodologies and real-world applications. She is currently an Applied Scientist at Amazon, within the Global Store organization, leading projects in E-Commerce Recommendation Systems, Search Engines, Multi-Modal Vision-Language Models (VLMs), and LLM/GenAI in retails. Her work has been published in top-tier conferences including ICCV, CVPR, ICLR, SIGIR, WACV, etc. At ICCV 2025, she won the Best Paper Award in Retail Vision. Spatial Mental Modeling from Limited Views Can VLMs imagine the unobservable space from just a few views, like humans do? Humans form spatial mental models, as internal representations of "unseen space" to reason about layout, perspective, and motion. On our proposed MINDCUBE, we see critical gap systematically on VLMs building robust spatial mental models through representing positions (cognitive mapping), orientations (perspective-taking), and dynamics (mental simulation for ''what-if'' movements). We then explore three approaches to help VLMs approximate spatial mental models, including unseen intermediate views, natural language reasoning chains, and cognitive maps. The significant improvement comes from ''map-then-reason'' that jointly trains the model to first abstract a cognitive map and then reason upon it. By training models to construct and reason over these internal maps, we boosted accuracy from 37.8% to 60.8% (+23.0%). Adding reinforcement learning pushed performance even further to 70.7% (+32.9%). Our key insight is that such scaffolding of spatial mental models, actively constructing and utilizing internal structured spatial representations with flexible reasoning processes, significantly improves understanding of "unobservable space". We aim to understand why geometric concepts remain challenging for VLMs and outlining promising research directions towards fostering more robust spatial intelligence. About the Speaker Manling Li is an Assistant Professor at Northwestern University and Amazon Scholar. She was a postdoc at Stanford University, and obtained the PhD degree in Computer Science at University of Illinois Urbana-Champaign in 2023. She works on the intersection of language, vision, and robotics, recognized by the MIT TR 35 Under 35, ACL Inaugural Dissertation Award Honorable Mention, ACL’24 Outstanding Paper Award, ACL'20 Best Demo Paper Award, and NAACL'21 Best Demo Paper Award, Microsoft Research PhD Fellowship, EE CS Rising Star, etc. Forecasting and Visualizing Air Pollution via Sky Images and VLM-Guided Generative Models Air pollution monitoring is traditionally limited by costly sensors and sparse data coverage. Our research introduces a vision-language model framework that predicts air quality directly from real-world sky images and also simulates skies under varying pollution levels to enhance interpretability and robustness. We further develop visualization techniques to make predictions more understandable for policymakers and the public. This talk will present our methodology, key findings, and implications for sustainable urban environments. About the Speaker Mohammad Saleh Vahdatpour is a PhD candidate in Computer Science at Georgia State University specializing in deep learning, vision–language models, and sustainable AI systems. His research bridges generative AI, environmental monitoring, and motion perception, focusing on scalable and energy-efficient models that connect scientific innovation with real-world impact. Sari Sandbox: A Virtual Retail Store Environment for Embodied AI Agents We present Sari Sandbox, a high-fidelity, photorealistic 3D retail store simulation for benchmarking embodied agents against human performance in shopping tasks. Addressing a gap in retail-specific sim environments for embodied agent training, Sari Sandbox features over 250 interactive grocery items across three store configurations, controlled via an API. It supports both virtual reality (VR) for human interaction and a vision language model (VLM)-powered embodied agent. We also introduce SariBench, a dataset of annotated human demonstrations across varied task difficulties. Our sandbox enables embodied agents to navigate, inspect, and manipulate retail items, providing baselines against human performance. We conclude with benchmarks, performance analysis, and recommendations for enhancing realism and scalability. About the Speakers Emmanuel G. Maminta is a fourth-year Artificial Intelligence Ph.D. student at the Ubiquitous Computing Laboratory (UCL) in the University of the Philippines Diliman, advised by Prof. Rowel O. Atienza. Janika Deborah B.Gajo is an undergraduate student studying for a Bachelor of Science in Computer Engineering at the University of the Philippines, Diliman. |
Nov 20 - Best of ICCV (Day 2)
|
|
Spatial Data Science Conference 2023 #SDSC23 London – Highlights Reel
2023-08-29 · 16:09
Did you miss The Spatial Data Science Conference in London 2023? No worries, you can see the highlights here. If you want to see more of the presentations from KFC, CARTO, Databricks, AWS, BT, Choreograph and Mott Macdonald - see our full content hub: https://spatial-data-science-conference.com/2023/london/ |
|
|
Vector and Raster Data Unification Through H3 | M. Colic | Tech Lead Public Sector UK&I | Databricks
2023-08-29 · 16:09
Milos Colic
– Tech Lead Public Sector UK&I
@ Databricks
Milos Colic, Tech Lead Public Sector UK&I at Databricks, demonstrates how raster and vector geospatial data can be standardised into a unified domain. This unification facilitates an easy plugin/plugout capability for all raster and vector layers. Databricks used these principles to design an easy, scalable and extensible Flood Risk for Physical Assets solution using H3 as a unification grid. To learn more about H3 check out: https://docs.carto.com/data-and-analysis/analytics-toolbox-for-bigquery/sql-reference/h3 |
|
|
Selecting EV Chargepoint at Connected Kerb: What We Know & What We Don't | A. Saad. & S. Knorr
2023-08-29 · 16:09
Ali Saad
– Lead Data Scientist
@ Connected Kerb
,
Susanne Knorr
– Product Owner Site Selection
@ Connected Kerb
Ali Saad, lead data scientist at Connected Kerb and Susanne Knorr, Product Owner Site Selection at Connected Kerb give an overview to Connected Kerb's journey experimenting to find the influencing factors and variables to predict high utilization of future charging point locations as well as the many questions that come with solving the challenge of accurate prediction in a highly dynamic environment. Learn more about EV Charger site selection here: https://carto.com/industries/utilities |
|
|
Improving Urban Mobility through Geospatial Analytics | Fawad A. Qureshi | Snowflake
2023-08-29 · 16:09
Fawad A. Qureshi
– Industry Field CTO
@ Snowflake
Fawad A. Qureshi, Industry Field CTO at Snowflake, explores how Voi Technologies, a Swedish e-scooter sharing company, is revolutionizing urban mobility through the power of geospatial analytics. He also discusses the challenges of urban transportation and how Voi is using data-driven insights to optimize scooter placement, improve safety, and create more efficient transportation networks. By harnessing the power of geospatial analytics, Voi is redefining urban mobility and improving the overall quality of life in cities. Join us to learn how data can transform the way we move around our cities. To learn more about mobility check out: https://carto.com/solutions/mobility-planning |
|
|
Understanding Customer Sentiment for Better LI : The Case of Campari | The Data Appeal Company
2023-08-29 · 16:09
Hannah Babineau
– Head of Partnerships
@ Data Appeal
,
Bernardo Monechi
– Scientist
@ The Data Appeal Company
In this presentation, Bernardo Monechi, Scientist at The Data Appeal Company and Hannah Babineau, Head of Partnerships at Data Appeal explore the value of Sentiment Analysis and the Feedback Economy. They reveal how global beverage brand, Campari, harnesses the power of online reviews to pinpoint the most strategic places to distribute their products and launch their brand in new markets. By combining sentiment analysis with points of interest data, discover how Campari expanded their aperitivo favorites into Istanbul, Turkey. To learn more about the CARTO x The Data Apeal Company tool check out our blog: https://carto.com/blog/data-appeal-company-poi-sentiment-data-available |
|
|
Emerging from Disruptive Times: The Changing Role of our Towns & Cities | Louise Etherden | CACI Ltd
2023-08-29 · 16:09
Louise Etherden
– Partner
@ CACI Ltd
After a turbulent few years and with the storm still raging, Louise Etherden, Partner at CACI Ltd., looks at how consumer behaviour is changing, what that has meant for the retail landscape, living and working patterns and what this means for our towns and cities. Check out CARTO behavioral data sets : https://carto.com/spatial-data-catalog/behavioral-data/ |
|
|
Complex Spatial Data Science in the Boardroom | Katy Ashwin & Blair Freebairn | KFC UK & Geolytix
2023-08-29 · 16:09
Katy Ashwin
– Marketing Planning Analyst
@ KFC UK
,
Blair Freebairn
– CEO
@ Geolytix
Where should we open 50 stores? Simple question right? To answer it Katy Ashwin, Marketing Planning Analyst at KFC UK and Blair Freebairn, CEO of Geolytix, talk through complex spatial modelling, as well as mobility data derived interaction surfaces, spatial ML ensemble models by channel and store format, big data optimization to create opportunity heat surfaces, and much more. Learn more about site selection : https://carto.com/solutions/site-selection |
|
|
Retail & Commerce : Driving Innovation with Location Intelligence | YemeTech & Dataplor
2023-08-29 · 16:08
Alejandro Quinto
– Head of Innovation
@ YemeTech
,
Dr. Beth Crane
– Chief Data Operations Officer
@ Dataplor
Dr. Beth Crane, Chief Data Operations Officer at Dataplor will be discussing the importance of location intelligence and how it can revolutionize your business. You'll also hear from Alejandro Quinto, Head of Innovation at YemeTech discuss YemeTech's implementation of location intelligence solutions and learn best practices and tips for doing it yourself. Discover the various ways POI data can be used to optimize store locations, understand consumer behavior, and make more informed decisions that can save costs and increase revenue. Learn more about location intelligence in retail here : https://carto.com/industries/retail |
|
|
Expert Panel | Sustainability
2023-08-29 · 16:08
Biswajit Acharya
– Ph.D., Consulting Partner, Sustainable Banking, Finance and Investment
@ Tata Consultancy Services
,
Jasmine Small
– Data Science and Research Officer
@ Marine Stewardship Council (MSC)
,
Dr Andrew Smith
– Co-Founder
@ Fathom
,
Caroline Robinson
– Unite Lead Executive Board
@ Women+ in Geospatial
This panel on sustainability features Caroline Robinson, Unite Lead Executive Board at Women+ in Geospatial; Biswajit Acharya, Ph.D. , Consulting Partner, Sustainable Banking, Finance and Investment at Tata Consultancy Services; Dr Andrew Smith, Co-Founder at Fathom; and Jasmine Small, Data Science and Research Officer at Marine Stewardship Council (MSC). They focus on the role of geospatial data and technology in sustainability, and how it can be best used to tackle challenges and help organizations reach their goals in this area. Learn more about CARTO's mission to deliver technology, data and scalability to help global businesses and communities make strategic decisions surrounding their sustainability programs. https://sustainability.carto.com/ |
|
|
Aaron Martin
– Data Innovation Manager
@ Clear Channel UK
,
Bryan Bonack
– Director of Product Management
@ SafeGraph
As the out of home advertising landscape evolves, so does the scope of this answer. Aaron Martin, Data Innovation Manager at Clear Channel UK and Bryan Bonack, Director of Product Management at SafeGraph, discuss how OOH media companies must consider the challenges of ensuring only real, valid places meet the refined definition to enable advertisers with better campaign planning insights. To learn more about out of home advertising check out our webinars: https://carto.com/webinars/ooh-advertising |
|
|
Using Spatial Data to Help Cities Deliver Positive Outcomes | Ed Parham | Space Syntax
2023-08-29 · 16:08
Ed Parham
– Director of Innovation and Design
@ Space Syntax
Ed Parham, Director of Innovation and Design at Space Syntax discusses how over time, seemingly unimportant daily activities, such as walking to work, can have profound impacts on outcomes such as health. Space Syntax combines and analyses the inter-relationships between built environment systems to understand, from the point of a person, how the hidden structures these create relate to emergent activities and outcomes in cities. For more on urban planning using spatial data check out: https://carto.com/industries/cities-government |
|
|
Why Do Some Retail Stores Perform Better Than Others? | T. Backlar & A. Outman | Echo Analytics
2023-08-29 · 16:08
Thea Backlar
– VP of Product
@ Echo Analytics
,
Alexandre Outman
– Product Manager
@ Echo Analytics
Join Thea Backlar, VP of Product at Echo Analytics and Alexandre Outman, Product Manager at Echo analytics for a thought-provoking session to discover how mobility data can be used to solve the most perplexing and complex business insights conundrums. Get an inside look at how a large European retailer solved the mystery of why some stores outperformed others. To learn more about site selection check out: https://carto.com/solutions/site-selection |
|
|
Evan Harwin
– Data Scientist
@ Mott MacDonald
Cutting costs and emissions using a tailor-made geospatial algorithm that uses machine learning to optimise inspections - a crucial part of managing distributed assets. Optimising the problem in both space and time to cut the cost of changing schedule down to virtually nothing. |
|