Cybersecurity will never be the same
RESEARCH
-
Microsoft Azure’s AI Architecture Insights with Mark Russinovich
By
–
Next week, Microsoft @Azure
’s @markrussinovich will join us to discuss Microsoft’s AI architecture and share insights on trending AI research and other advances in the field. In-person attendance is now at capacity, but you can still join us via Zoom: -
Open Source Robotics Future: Fine-Tuning Robots Over LLMs
By
–
I can see that, and I agree. It may take a bit of time to catch up, but open source will get there. I dream about a future where we fine-tune robots instead of LLMs, and I am sure we'll get there.
-
AI as Automated Statistical Inference: Definition and Scope
By
–
AI simply means automated statistical inference here. The algorithms, data and setup have very little in common beyond this.
-
Advanced RAG Series: Indexing Optimization Techniques
By
–
Advanced RAG series: Indexing Indexing is a huge part or RAG. Lot's of different stuff to optimize: 1. Chunk Optimization
2. Multi-representation indexing
3. Specialized Embeddings
4. Hierarchical Indexing Great article by @divyanshu_van diving into all these topics! Blog: -
Trustworthy AI Principles Guide Technology Development
By
–
Learn about the #trustworthyAI principles that guide our technology development, partner solutions, and research and methodologies.
-
Evaluating Resource Consumption and Value Proportionality Beyond Size
By
–
Yes I did, and I think it’s a useful comparison. I also think it’s good to evaluate resource consumption beyond size comparisons and consider whether the value it brings is proportional
-
LLMs Medical Claims Citation Accuracy Under Research Scrutiny
By
–
Researchers are studying how often and how accurately LLMs cite references when making medical claims – and finding room for improvement.
-
BM25 Search Implementation Performance Comparison with ElasticSearch
By
–
also shout out to @clifapt for running BM25 via ES on LoCo first: https://
x.com/clifapt/status
/1755303087873429667
… my implementation seems to be slightly less performant than ElasticSearch; not sure why. maybe because i'm using a subword tokenizer, or suboptimal hyperparams? -
Fast GPU-Enabled BM25 Implementation in PyTorch Achieves SOTA
By
–
implemented a fast, GPU-enabled BM25 in pytorch! BM25 is a simple search algorithm from the 70s that works as well as neural networks for most search problems; for all the advances we've made in neural text retrieval, it's still around got near SOTA on stanford LoCO benchmark