Paper Summary β OSWorld-Human
This is an independent score computing method as a supplement to OSWorld. In this work, the authors design a brand new metric to evaluate agent's work efficiency.
From Neural Network to AI
This document records the main procedure of AI development, from neural network to CNN to RNN to Transformer to LLM to agent...
Typora Guide
Too lazy to describe.
How to Use Personal Knowledge Graph
A practical guide to my personal knowledge graph: notes, edges, evidence, queries, and how it connects to this blog.
Notes on Machine Learning and Deep Learning and Reinforcement Learning
Theory of Machine Learning & Deep Learning & Reinforcement Learning
Mechanism of Agent
Describe the mechanism of Agent architecture.
SWE-agent Traj Analysis
A report of a traj file of SWE-agent.
Paper Summary β NSI
Paper Summary of "Lifting Traces to Logic: Programmatic Skill Induction with Neuro-Symbolic Learning for Long-Horizon Agentic Tasks".