§ BEAT
Research
TFlow cuts multi-agent inference tokens 83% via weight injection
IBM Boosts Zero-Shot Search Accuracy 25% With LLM Query Refinement
Optimizer-Model Consistency Cuts LLM Forgetting in Finetuning
Sparse MoEs retain accuracy at 87.5% weight pruning
LongSeeker Beats Competitors on Long-Horizon Tasks
OpenSeeker-v2 beats Alibaba's Tongyi on agentic search benchmarks
Automated agent recommender cuts multi-agent system engineering steps to one
Knowledge Distillation Brings Code Clone Detection On-Premise
Diffusion Models Cut Compute on Sparse Data with Selective Processing
RunAgent Enforces Deterministic Execution for LLM Workflows
Columbia Releases Open-Source 100 Hz Tactile Sensor
Microsoft generates 1,000 synthetic computers to train agents
Carbon-Taxed Transformers Cut Model Memory 49x Without Retraining
Alec Radford Releases 13B Model Trained on Pre-1931 Text Under Apache 2.0
Tencent Open-Sources HunyuanWorld 1.0, a Mesh-Ready 3D World Generator
GiVA cuts vector fine-tuning rank 8-fold to match LoRA training speed
