AI Research

A new arXiv paper (2609.16073v1) formalizes "State Mutability" in Retrieval-Augmented Generation memory for long-horizon autonomous agents, proving that standard dense retrieval suffers from Asymptotic Recall Decay and t…

Researchers proposed RAG-CT, a defense that detects malicious queries in Retrieval-Augmented Generation (RAG) systems by analyzing their entropy and margin distributions with a score-based detection method, according to …

A new arXiv paper (2609.16056v1) formalizes behavioral knowledge as a precondition Bayesian network over an agent's structural actions and tests three injection points in a neuro-symbolic RL loop: a symbolic verifier use…

Researchers proposed Causal Neural Set Filtering (CNSF), a neural set filter that encodes only current measurements while carrying past evidence in a structured recursive track state, reducing mean GOSPA and T-GOSPA rela…

A two-stage modeling framework combining a binary maneuver-decision predictor with a decision-conditioned autoregressive Transformer achieved 0.49 m/s² acceleration MAE and 0.62 m distance MAE in predicting driver behavi…

A paper submitted to arXiv on 13 Sep 2026 introduces signed, weighted affine p-adic residual objectives as native encodings of finite-domain constraints, with a coordinatewise domination theorem guaranteeing every global…

Researchers introduced HintMiner, an automatic question-hints mining tool that extracts hints from web Q&A posts using MiningNet, a transformer-based neural encoder-decoder model with copying mechanisms trained via a sel…

A transformer-enhanced neural operator (TNO) predicted transonic compressor blade performance with lower error than state-of-the-art deep learning operators FNO and DeepONet, according to an arXiv paper (2609.16066v1) tr…

A new arXiv paper, arXiv:2609.16065v1, introduces Agentic Heuristic Learning (AHL) Studio, a tool that performs human activity recognition (HAR) without gradient-based neural network training. The tool uses a learning-ti…

Researchers proposed a semantics-consistent tabular diffusion framework, described in arXiv paper 2609.16069v1, that uses large language models to extract intra-column semantics and inter-column symbolic rules from metad…

A new arXiv paper (2609.16155v1) presents a framework that fine-tunes pre-trained Large Language Models and applies Retrieval Augmented Generation (RAG) to generate synthetic time series data for manufacturing processes,…

A schema-adaptive action-conditioned Joint-Embedding Predictive Architecture (SAAC-JEPA) for CNC dynamics achieved zero-shot RMSE of 0.546, R² of 0.012, and NLL of 0.52 on a sealed target machine sharing only 10 of the s…

A new arXiv paper (2609.16161v1) presents an end-to-end integer-only quantization approach and a sparse dictionary-based KV cache compression strategy that enables LLM inference on Flash compute-in-memory devices, reduci…

A new parameter-pruning scheme derived from the Fisher information metric's geodesic distance in model space outperforms magnitude pruning and local Fisher information pruning across every architecture and dataset combin…

A new arXiv review (2609.16213v1) finds that AI uplift for biological threats currently exists primarily in digital tasks, with frontier systems exceeding expert baselines on in-silico and screening-evasion benchmarks, w…

A new arXiv paper (2609.16245v1) introduces Metacognitive Steering, an inference-time controller that reads the cognitive regime of Kimi 2.6, a trillion-parameter mixture-of-experts model, and dynamically composes layer-…

A corpus of 2,507 head-to-head comparisons across 27 scientific disciplines from papers published between 2000 and early 2025 shows AI is not a universal replacement for existing scientific methods, according to a new ar…

Researchers proposed CLEAR, an agentic framework for cross-source evidence adjudication in large language models in medicine, detailed in arXiv paper 2609.16301v1. CLEAR independently generates candidate answers from thr…

Researchers introduced RAIL, an interface exposing polyhedral enclosures of nonlinear dynamics to LiRPA-style bound propagation, and CLIPPER, a branch-and-bound algorithm that jointly refines enclosures and splits contro…

Researchers introduced Blindspot, a benchmark for trajectory-level safety calibration of long-horizon tool-using LLM agents, containing 22 attack families and 35 scenarios across seven domains that yield more than 2,500 …