Author: Sajjad Ansari

Sajjad Ansari
292 POSTS0 COMMENTS
Sajjad Ansari is a final year undergraduate from IIT Kharagpur. As a Tech enthusiast, he delves into the practical applications of AI with a focus on understanding the impact of AI technologies and their real-world implications. He aims to articulate complex AI concepts in a clear and accessible manner.

From Fixed to Random Designs: Unveiling the Hidden Factor Behind Modern Machine Learning ML Phenomena

Modern machine learning (ML) phenomena such as double descent and benign overfitting have challenged long-standing statistical intuitions, confusing many classically trained statisticians. These phenomena...

Block Transformer: Enhancing Inference Efficiency in Large Language Models Through Hierarchical Global-to-Local Modeling

Large language models (LLMs) have gained widespread popularity, but their token generation process is computationally expensive due to the self-attention mechanism. This mechanism requires...

Self-Training on Image Comprehension (STIC): A Novel Self-Training Approach Designed to Enhance the Image Comprehension Capabilities of Large Vision Language Models (LVLMs)

Large language models (LLMs) have gained significant attention due to their advanced capabilities in processing and generating text. However, the increasing demand for multimodal...

Unraveling Transformer Optimization: A Hessian-Based Explanation for Adam’s Superiority over SGD

Large Language Models (LLMs) based on Transformer architectures have revolutionized AI development. However, the complexity of their training process remains poorly understood. A significant...

Revisiting Weight Decay: Beyond Regularization in Modern Deep Learning

Weight decay and ℓ2 regularization are crucial in machine learning, especially in limiting network capacity and reducing irrelevant weight components. These techniques align with...

A Novel AI Approach to Multicut-Mimicking Networks for Hypergraphs with Constraints

Graph sparsification is a fundamental tool in theoretical computer science that helps to reduce the size of a graph without losing key properties. Although...

Simplifying Diffusion Models: Fine-Tuning for Faster and More Accurate Depth Estimation

Monocular depth estimation (MDE) plays an important role in various applications, including image and video editing, scene reconstruction, novel view synthesis, and robotic navigation....

Chain-of-Thought (CoT) Prompting: A Comprehensive Analysis Reveals Limited Effectiveness Beyond Math and Symbolic Reasoning

Chain-of-thought (CoT) prompting has emerged as a popular technique to enhance large language models’ (LLMs) problem-solving abilities by generating intermediate steps. Despite its better...

LoRID: A Breakthrough Low-Rank Iterative Diffusion Method for Adversarial Noise Removal

Neural networks are widely adopted in various fields due to their ability to model complex patterns and relationships. However, they face a critical vulnerability...

SynSUM: A Synthetic Benchmark for Integrating Clinical Notes with Structured Data

Electronic Health Records (EHRs) present a wealth of information, combining structured tabular data and unstructured clinical notes. This valuable resource forms the foundation for...

A Systematic Literature Review: Optimization and Acceleration Techniques for LLMs

Large language models (LLMs) have seen remarkable success in natural language processing (NLP). Large-scale deep learning models, especially transformer-based architectures, have grown exponentially in...

CONClave: Enhancing Security and Trust in Cooperative Autonomous Vehicle Networks Cooperative Infrastructure Sensors Environments

The cooperative operation of autonomous vehicles can greatly improve road safety and efficiency. However, securing these systems against unauthorized participants poses a significant challenge....