MRGSEM-Sum: An Unsupervised Multi-document Summarization Framework based on Multi-Relational Graphs and Structural Entropy Minimization
arxiv.org·1d
📈Graph Algorithms
RedCoder: Automated Multi-Turn Red Teaming for Code LLMs
arxiv.org·2d
🤖AI
Executing arbitrary Python code from a comment
hacktron.ai·4d·
Discuss: Lobsters, Hacker News
🎭Proc Macros
Concrete Security Bounds for Simulation-Based Proofs of Multi-Party Computation Protocols
arxiv.org·2d
🧠Memory Allocators
Cell-Free Massive MIMO SWIPT with Beyond Diagonal Reconfigurable Intelligent Surfaces
arxiv.org·1d
🧮Linear Algebra
Invisible Architectures of Thought: Toward a New Science of AI as Cognitive Infrastructure
arxiv.org·1d
🤖AI
Building an AI-native multi-model UI with SurrealDB
surrealdb.com·5d·
Discuss: Hacker News
🗄️Database Internals
Many prediction markets would be better off as batched auctions
lesswrong.com·15h
⏱️Real-time Systems
Does Editing Improve Answer Quality on Stack Overflow? A Data-Driven Investigation
arxiv.org·3d
🏗️LLVM
Research Areas in AI Control (The Alignment Project by UK AISI)
lesswrong.com·1d
🤖AI
Design of a bioinspired robophysical antenna for insect-scale tactile perception and navigation
arxiv.org·1d
🎮Game Engines
GlideinBenchmark: collecting resource information to optimize provisioning
arxiv.org·3d
🧠Memory Allocators
Ethersync: Peer-to-peer collaborative editing of local text files
github.com·4d·
Discuss: Lobsters, Hacker News
📁File Systems
Interacting With Text Adventures Through Perl
entropicthoughts.com·5d·
Discuss: Lobsters, Hacker News
🐪Perl
Using Scaling Laws for Data Source Utility Estimation in Domain-Specific Pre-Training
arxiv.org·2d
🕯️Candle
Snap, Segment, Deploy: A Visual Data and Detection Pipeline for Wearable Industrial Assistants
arxiv.org·3d
🎮Game Engines
Discrete Tokenization for Multimodal LLMs: A Comprehensive Survey
arxiv.org·1d
🗣️Language Servers
The Observer Effect for belief measurement
lesswrong.com·13h
🌍Distributed Systems
LENS: Learning Ensemble Confidence from Neural States for Multi-LLM Answer Integration
arxiv.org·1d
🕯️Candle
DICOM De-Identification via Hybrid AI and Rule-Based Framework for Scalable, Uncertainty-Aware Redaction
arxiv.org·1d
🤖AI