Jonas Rothfuss's Projects
Deep neural network architecture for representing robot experiences in an episodic-like memory which facilitates encoding, recalling, and predicting action experience - Research Project at KIT's High Performance Humanoids Technologies Lab (H2T)
Web app that mines a user's bank statement history to suggest vouchers and coupons
A job launching library for docker, EC2, etc.
Improved Fisher Vector Implementation- extracts Fisher Vector features from your data
Generationenkontakt App
Tree-based representation of news articles based on Microsoft Azure Cognitive Services - Prototype created during the HackX 2017 Hackathon (Microsoft, Handelsblatt, FlowFact)
Sentiment analysis on stock related Tweets and subsequent investigation of investor sentiment impact on stock prices
Meta-learning Gaussian process (GP) priors via PAC-Bayes bounds
A logger, server and visualization dashboard for ML projects
Implementation of the Model-Based Meta-Policy-Optimization (MB-MPO) algorithm
Implementation of the paper "Movement Primitives via Optimization" (Dragan et al., 2016). It includes both the adaptation of trajectories with DMP and learning a better adaptation norm.
Implementation of Proximal Meta-Policy Search (ProMP) as well as related Meta-RL algorithm. Includes a useful experiment framework for Meta-RL.
Python wrapper for Stanford CoreNLP
playground repo for rl algorithms
rllab is a framework for developing and evaluating reinforcement learning algorithms, fully compatible with OpenAI Gym.
Skript zur Vorlesung "Sicherheit" am KIT.
tensorflow introductory presentation with example code for a small workshop at the H2T (KIT)
Texas Hold'em Poker Game developed with AngularJS, Pusher, Ruby on Rails & MongoDB
Easily convert RGB video data (e.g. .avi) to the TensorFlow tfrecords file format for training e.g. a NN in TensorFlow. This implementation allows to limit the number of frames per video to be stored in the tfrecords.
A Pipline for extracting and processing features from videos
Starter code for working with the YouTube-8M dataset.