talk-data.com talk-data.com

Event

Data Science Retreat Demo Day #35

2023-09-19 – 2023-09-19 Meetup Visit website ↗

Activities tracked

5

Data Science Retreat presents 5 Machine Learning prototypes and projects by Batch 35 participants. The event is free to attend and thanks to Sopra Steria SE for hosting us.

Agenda:

17:30 - Drinks and Networking

18:00 - Welcome & Introduction

Followed by Project Presentations

PROJECT IDEAS:

1. Social media Stats Servant (Social media debugging) Project by: Daniele Didino and Marco Zausch

The idea is to evaluate a text (e.g., a Tweet) and to provide references pointing to an official source (e.g., official statistics from the German federal government). Since automatic fact checking proves to be hard and a slippery affair, the stats servant is supposed to make it easy for humans to spot and counter disinformation. By delivering official statistics in reference to allegations made on social media, we build a little helper to sober debate on the internet. Starting small with statistics for a narrowly defined subject area it would be nice to branch out to areas populists usually exploit.

2. Value Investing Hypothesis testing Project by: Emre Oymakapu

Testing Benjamin Graham`s value investing system in stock market using selected value investing features of data input from the quarterly financials of companies. Value investing is interested in the real calculated value of each company according to its financials and has no interest in the price fluctuations of the market. Hence a portfolio of stocks are selected that are ‘cheap’ and long term return are collected until the stock price reachs saturation. Project is going to feed in these features (collected from historical stock data) into a model and hope to come up with a better weighted system with the help of data science and machine learning practices.

3. AI based Wind Power Predictions Project by: Oliver Eisele

In recent years, and with increasing installation of renewable energy, the demand for accurate wind power prediction is growing, in order to keep grid stability at a high level, This project aims to use rather large scale numerical weather predictions and map it to local measurements. In a next step local weather predictions are combined with historical operational data of a large wind farm and the production in the next 48h is forecasted.

4. CodeWhiz- the specialized coding assistant Project by: Malkhaz Pertaia and Telma Regi

We are embarking on the creation of a specialized programming assistant, meticulously fine-tuned for libraries such as PyTorch, TensorFlow, or Dart, FastAIs. This intelligent assistant, accessible through a chat-like interface, is designed to offer tailored guidance, provide access to the latest documentation, and suggest learning resources to users. It will comprehend contextual queries, ensuring deep library expertise, and offer direct links to official documentation, facilitating efficient problem-solving and learning. With continuous updates, personalization options, and a commitment to privacy, this coding assistant aims to significantly enhance the development experience for programmers and serve as an invaluable resource in the ever-evolving landscape of software development.

5. Teach papers to 5 years old Project by: Anahita Pakiman and Jonathan Abadie

Wanna be able to read a paper from a domain you know nothing about ? Use our tool to get explanations of every term used in the paper. A concept is used in the explanation that you don’t know ? Go deeper again until you reach the explanation of for example addition or any simple concept that a 5 years old could understand, if he knows how to read of course ! Built using ChatGPT and other sources.

See you all at the event.

Sessions & talks

Showing 1–5 of 5 · Newest first

Search within this event →

AI based Wind Power Predictions

2023-09-19
talk
numerical weather predictions wind farm data forecasting

In recent years, and with increasing installation of renewable energy, the demand for accurate wind power prediction is growing, in order to keep grid stability at a high level, This project aims to use rather large scale numerical weather predictions and map it to local measurements. In a next step local weather predictions are combined with historical operational data of a large wind farm and the production in the next 48h is forecasted.

CodeWhiz- the specialized coding assistant

2023-09-19
talk
PyTorch TensorFlow dart fastai chatbot documentation

We are embarking on the creation of a specialized programming assistant, meticulously fine-tuned for libraries such as PyTorch, TensorFlow, or Dart, FastAIs. This intelligent assistant, accessible through a chat-like interface, is designed to offer tailored guidance, provide access to the latest documentation, and suggest learning resources to users. It will comprehend contextual queries, ensuring deep library expertise, and offer direct links to official documentation, facilitating efficient problem-solving and learning. With continuous updates, personalization options, and a commitment to privacy, this coding assistant aims to significantly enhance the development experience for programmers and serve as an invaluable resource in the ever-evolving landscape of software development.

Social media Stats Servant (Social media debugging)

2023-09-19
talk
text analysis NLP

The idea is to evaluate a text (e.g., a Tweet) and to provide references pointing to an official source (e.g., official statistics from the German federal government). Since automatic fact checking proves to be hard and a slippery affair, the stats servant is supposed to make it easy for humans to spot and counter disinformation. By delivering official statistics in reference to allegations made on social media, we build a little helper to sober debate on the internet. Starting small with statistics for a narrowly defined subject area it would be nice to branch out to areas populists usually exploit.

Teach papers to 5 years old

2023-09-19
talk
explanation generation llms educational tooling

Wanna be able to read a paper from a domain you know nothing about ? Use our tool to get explanations of every term used in the paper. A concept is used in the explanation that you don’t know ? Go deeper again until you reach the explanation of for example addition or any simple concept that a 5 years old could understand, if he knows how to read of course ! Built using ChatGPT and other sources.

Value Investing Hypothesis testing

2023-09-19
talk
stock market data machine learning

Testing Benjamin Graham's value investing system in stock market using selected value investing features of data input from the quarterly financials of companies. Value investing is interested in the real calculated value of each company according to its financials and has no interest in the price fluctuations of the market. Hence a portfolio of stocks are selected that are cheap and long term return are collected until the stock price reachs saturation. Project is going to feed in these features (collected from historical stock data) into a model and hope to come up with a better weighted system with the help of data science and machine learning practices.