The Null Hypothesis

Experiments, Thoughts & Research on
`Multimodal Learning, Inference Optimisation Computer Vision and Deep Learning`

Object-Guided Visual Tokens:
Eliciting Compositional Reasoning
in Multimodal Language Models

Addressing MLLMs shortcomings in Compositional Reasoning through CLIP-Segmentation fusion

9 min read · September 05, 2025

2025 · Multimodal-Learning Compositional-Reasoning Computer-Vision
Perception, Localization, Planning and Control on RAE Robots

Computer Vision for Autonomour Robots

17 min read · January 20, 2025

2025 · Computer-Vision Robotics
Model Compression for Machine Translation in Large Language Models

Model Compression for Machine Translation in Large Language Models

7 min read · November 05, 2024

2024 · Inference-Optimization LLMs
Optimizing Predictions: Vocabulary Reduction and Contrastive Decoding in LLMs

Efficiency-focused early exiting, vocabulary pruning, and contrastive decoding for LLM inference

43 min read · October 26, 2024

2024 · Inference-Optimization LLMs Early-Exiting Contrastive-Decoding Efficiency