Name: Gerard Planella
Type: User
Company: Universiteit van Amsterdam
Bio: Artificial Intelligence MSc Student. Passionate about Geometric Deep Learning, Language Models and Generative Models.
Location: Amsterdam, The Netherlands
Gerard Planella's Projects
Repository for a Delivery Planning project, where we must find optimal delivery routes giving any input graph. The development was performed with ASP using the Clingo Python package for the Knowledge Representation and Reasoning subject for the Artificial Intelligence MSc at the University of Amsterdam 2023.
Implementation of the HomeSound Robot system
Sentence embeddings (InferSent) and training code for NLI.
Investigating the extent to which structural encodings in geometric methods contribute in capturing topological information and developing a generic framework (ToGePi) to inject topological and geometric information into MPNN architectures.
Code for a multi-agent particle environment used in the paper "Multi-Agent Actor-Critic for Mixed Cooperative-Competitive Environments"
Reproducibility and extension study for the Fairness for Cooperative Multi-Agent Learning with Equivariant Policies Reproducibility Research Paper
Study on stereotype transfer accross Multilingual Language Models for English, Spanish, French, Greek and Croatian. The emotion profiles for different social groups are obtained from pre-trained XLM-RoBERTa and fine-tuned versions of it.
A repository for explaining feature attributions and feature interactions in deep neural networks.
Development of a simple Classifier based on a Quantum Neural Network implemented with Pytorch and PennyLane
Implementation of different Sentence Encoders trained on Natural Language Inference and evaluated on unseen tasks through Meta's SentEval framework.
A python tool for evaluating the quality of sentence embeddings.
Implementation of Robotic Platform for Plant care and management
Analysis of soft-DMAs on low-cost ZYNQ devices for IoT platforms running bare-metal applications.