@mayhewsw: New paper: I know the hotness is all in 10xing compute scale, and telling things to think step by step with tool use, b…
Summary
Authors release Universal NER v2, a named-entity recognition paper presented at LREC 2026 that deliberately eschews modern scaling and tool-use trends.
Similar Articles
@dair_ai: NEW paper from Meta: Agentic Discovery of Neural Architectures. This is a hot new area of research! Keep an eye on it.
Meta's new paper presents an agentic system that autonomously discovers neural architectures outperforming Llama 3.2 at 350M, 1B, and 3B scales within a 24-hour compute budget.
@dair_ai: https://x.com/dair_ai/status/2056018543850754283
A roundup of the top AI papers from May 11-17, covering Lighthouse Attention for long-context pretraining, a comparison of grep vs embedding retrieval for coding agents, and mechanistic interpretability work revealing a geometric calculator in LLMs.
A Mechanism and Optimization Study on the Impact of Information Density on User-Generated Content Named Entity Recognition
ArXiv preprint identifies low information density as the root cause of NER performance collapse on noisy user-generated content and introduces the Window-Aware Optimization Module (WOM) that boosts F1 by up to 4.5% on WNUT2017.
GLiNER-Relex: A Unified Framework for Joint Named Entity Recognition and Relation Extraction
GLiNER-Relex is a unified framework for joint named entity recognition and relation extraction that leverages a shared transformer encoder for zero-shot capabilities. The paper demonstrates competitive performance on standard benchmarks and releases the model as an open-source Python package.
@dair_ai: NEW paper worth reading. GPT-5.4 nano plus a critic-comparator orchestration loop hits 76.4% on SWE-bench Verified, mat…
A new paper shows that using a weak model with k=8 proposals and a critic-comparator selection loop can match frontier model performance on SWE-bench Verified, reaching 76.4% accuracy. The key insight is that correct patches are often already present in a weak model's top-k candidates, and the challenge is effective selection using execution verification.