Harsh Trivedi

A Glimpse at Their Relationship and Design The traditional neural networks were not designed to handle the sequential dependencies of inputs. The Problem. If you go back in time, there was a brief period where researchers were trying to figure out newer architectures to solve the problem of neural nets not being able to understand the dependency of one input on the previous input. Most architectures were built on the assumption of processing the inputs at an independent level and not retain any context a...

Causal Inferencing : Finding the “cause” and the “effect” A blog about causality, discovery and estimation. Causality plays an important role when we try to understand a particular situation involving multiple variables that might be related and have an effect on each other. This blog mainly focuses on : Defining Causality The 3 Levels of Causal Reasoning Importance of DAG in causal inference Causal Algorithms. Causal tooling and implementations Defining Causality. The term “causality” is closely relat...

MCP 101 : All about Model Context Protocol. A blog about Model Context Protocol. So if you’ve been in touch with the whole “AI side” of the world or just been in touch with tech in general, I’m sure you must’ve heard about Anthropic releasing MCP in late 2024. Okay so lets dive a little deeper into MCP. What is MCP ? Model Context Protocol is basically a protocol that lets your LLMs access tools and information in a standard manner. Giving tool access to LLMs along with resources makes the LLM a much mo...