Kim 2 ML
Applied AI notes for data scientists, analysts, and builders
From AI papers to decisions you can apply this week.
I break down new AI and data science research into practical frameworks, implementation notes, and trade-offs that matter in production.
- Paper summaries with practical implementation takeaways
- Opinionated perspectives on what works in real projects
- Clear prompts for discussion with data science practitioners
Recent Posts
- Coverage – Aware Web Crawling for Domain – Specific Supplier Discovery via a Web – Knowledge – Web Pipeline
ref: https://arxiv.org/pdf/2602.24262 Summary Most knowledge graphs fail in production because they are incomplete and hard to maintain. This… Read more: Coverage – Aware Web Crawling for Domain – Specific Supplier Discovery via a Web – Knowledge – Web Pipeline - How to Train Your Deep Research Agent? Prompt, Reward, and Policy Optimization in Search-R1ref: https://arxiv.org/pdf/2602.19526v1 Summary I used to think longer prompts and step-by-step reasoning always make AI smarter. In daily… Read more: How to Train Your Deep Research Agent? Prompt, Reward, and Policy Optimization in Search-R1
- Prompt Repetition Improves Non-Reasoning LLMs
ref: https://arxiv.org/pdf/2512.14982 Summary The easiest AI upgrade might be just “copy and paste”. Google researchers tested with 7… Read more: Prompt Repetition Improves Non-Reasoning LLMs - Verbalized Sampling: How to mitigate Mode Collapse and unlock LLM diversityref: https://arxiv.org/pdf/2510.01171 Summary AI models repeat safe answers, but a simple prompt tweak can unlock far better thinking.… Read more: Verbalized Sampling: How to mitigate Mode Collapse and unlock LLM diversity
- Tool-MAD: A Multi-Agent Debate Framework for Fact Verification with Diverse Tool Augmentation and Adaptive Retrieval
ref: https://arxiv.org/pdf/2601.04742 Summary The team studied several methods to improve LLM accuracy and combined the strongest ideas to… Read more: Tool-MAD: A Multi-Agent Debate Framework for Fact Verification with Diverse Tool Augmentation and Adaptive Retrieval
Kim 2 ML 