Characterisation of Goalkeeper Actions from Skeletal Data
Characterised goalkeeper actions from skeletal data by learning robust representations via contrastive learning and analysing their structure using clustering techniques. Click to read more!
AI
Mechanistic Interpretability of VLM on Spatial Relational Reasoning
Studied LLaVA-1.5-7B using a controlled synthetic benchmark; revealed an existence bias causing yes-bias in binary relational tasks through logit lens, linear probing, and attention analyses, showing mid-layer representations are linearly decodable but misaligned with final decisions. Click to read more!
Evaluating and Finetuning LLMs for Multilingual Legal Summarisation
Adapted Apertus 8B/70B for Swiss legal summarisation, showing full fine-tuning outperforms GPT-4o and Claude 3.5 Sonnet on BERTScore and ROUGE, while highlighting trade-offs between LoRA and full fine-tuning. Click to read more!
Diffusion Models: An Intuitive Understanding
Disclaimer: This post is my notes on understanding diffusion models from an intuitive perspective. It is not a formal explanation, and I might have made mistakes. Please reach out to me if you find any errors! Introduction The original paper was from UC Berkley and went by the name Denoising Diffusion Probabilistic Model. The original idea might sound counter-intuitive, but it takes a random noise and transforms it into a realistic image step-by-step. ...