2026-09-14

Test temporal consistency and motion realism in video generation, Make RAG retrieval and knowledge-base QA more reliable, Make agents use tools and reusable skills more reliably

Today tracks: test temporal consistency and motion realism in video generation, make RAG retrieval and knowledge-base QA more reliable, make agents use tools and reusable skills more reliably.

View sources 中文

Published 2026-09-14 Target source 2026-09-12 Actual source 2026-09-10 Candidates 100 Featured 6 Tracked 20 Source date fallback

This issue fetched and deduplicated 100 candidate papers from the 2026-09-10 source date, then selected 6 featured papers and 20 additional mentions.

Featured

1Biology-in-the-loop: Amortized Adaptive Hit Discovery in CRISPR Screens🔗
2The Last AI Built by Humans: Toward Genuine Recursive Self-Improvement🔗
3Learning Agent-based Model Predictive Control for Holistic Vehicle Performance🔗
4Truncated Noisy Best-Response Algorithms: Toward Game Theoretic Learning with Safety Guarantees🔗
5Why Does Post-Training Quantization Work?🔗
6Generalization Analysis of Distributed Kernel-based Robust Gradient Descent Algorithms🔗

What is worth tracking today

Today’s high-signal papers point to: test temporal consistency and motion realism in video generation, make RAG retrieval and knowledge-base QA more reliable, make agents use tools and reusable skills more reliably. The notes below focus on the core problem, method signal, main claim, and keywords for each featured paper.

Featured papers: core problem, method signal, and keywords

test temporal consistency and motion realism in video generation

Biology-in-the-loop: Amortized Adaptive Hit Discovery in CRISPR Screens (Carl Edwards, Edward De Brouwer, Xiner Li, Namkyeong Lee, Ehsan Hajiramezanali, Anne Biton, et al.) 2609.11877 PDF

Signalthis paper targets the concrete research problem behind test temporal consistency and motion realism in video generation. It uses the title, abstract, and public signals around evaluation, benchmark, post-training, eval to frame the benchmarks and evaluation task, data, or evaluation flow to improve test temporal consistency and motion realism in video generation. The main claim is the title, abstract, and public signals indicate: Many biological discovery problems require experiments to be selected sequentially under constrained budgets

Keywordsevaluationbenchmarkpost-trainingeval

Code/DataCheck the source paper

make RAG retrieval and knowledge-base QA more reliable

The Last AI Built by Humans: Toward Genuine Recursive Self-Improvement (Yi Duan, Ying Liu, Zirui Tang, Haodong Chen, Jun Zhou, Yumou Liu, et al.) 2609.11873 PDF

Signalthis paper targets the concrete research problem behind make RAG retrieval and knowledge-base QA more reliable. It uses the title, abstract, and public signals around search, index, rag, Retrieval and RAG to frame the retrieval and rag task, data, or evaluation flow to improve make RAG retrieval and knowledge-base QA more reliable. The main claim is the title, abstract, and public signals indicate: Recursive self-improvement (RSI) enables AI systems to turn experience and feedback into persistent changes that improve both their capabilities and the process of future improvement

KeywordssearchindexragRetrieval and RAG

Code/DataCheck the source paper

make agents use tools and reusable skills more reliably

Learning Agent-based Model Predictive Control for Holistic Vehicle Performance (Jiaming Zhong, Reza Valiollahi Mehrizi, Mohammad Pirani, Chao Yu, Alireza Kasaiezadeh, Yash Vardhan Pant, et al.) 2609.11871 PDF

Signalthis paper targets the concrete research problem behind make agents use tools and reusable skills more reliably. It uses the title, abstract, and public signals around agent, rag, safety, agents to frame the agents and tool use task, data, or evaluation flow to improve make agents use tools and reusable skills more reliably. The main claim is the title, abstract, and public signals indicate: Agent-based model predictive control (AMPC) has recently been proposed as a distributed scheme that collaborates with all agents to achieve optimal holistic performance

Keywordsagentragsafetyagents

Code/DataCheck the source paper

make agents use tools and reusable skills more reliably

Truncated Noisy Best-Response Algorithms: Toward Game Theoretic Learning with Safety Guarantees (Vartika Singh, Philip N. Brown) 2609.11863 PDF

Signalthis paper targets the concrete research problem behind make agents use tools and reusable skills more reliably. It uses the title, abstract, and public signals around agent, safety, agents, Agents and Tool Use to frame the agents and tool use task, data, or evaluation flow to improve make agents use tools and reusable skills more reliably. The main claim is the title, abstract, and public signals indicate: We consider a game theoretic approach to solve multi-agent coordination problems with submodular maximization objectives

KeywordsagentsafetyagentsAgents and Tool Use

Code/DataCheck the source paper

improve code generation, execution feedback, and automated repair

Why Does Post-Training Quantization Work? (Yuxiang Chen, Michael Beyer, Jun Zhu, Jianfei Chen) 2609.11716 PDF

Signalthis paper targets the concrete research problem behind improve code generation, execution feedback, and automated repair. It uses the title, abstract, and public signals around post-training, training, Training and Post-training to frame the training and post-training task, data, or evaluation flow to improve improve code generation, execution feedback, and automated repair. The main claim is the title, abstract, and public signals indicate: Post-training quantization compresses large language models (LLMs) by storing their weights at reduced precision, and each quantized weight introduces an error into the hidden states

Keywordspost-trainingtrainingTraining and Post-training

Code/DataCheck the source paper

track a high-signal other paper

Generalization Analysis of Distributed Kernel-based Robust Gradient Descent Algorithms (Jun-Yi Meng, Zheng-Chu Guo, Yuan Mao) 2609.11712 PDF

Signalthis paper targets the concrete research problem behind track a high-signal other paper. It uses the title, abstract, and public signals around other to frame the other task, data, or evaluation flow to improve track a high-signal other paper. The main claim is the title, abstract, and public signals indicate: In this paper, we investigate the generalization performance of distributed gradient descent algorithms in a reproducing kernel Hilbert space under a robust loss function $l_σ$

Keywordsother

Code/DataCheck the source paper

Other papers worth tracking

Beyond Word Error Rate: A Switch Aware Evaluation of ASR and Audio Language Models on English Yoruba Code-Switched Speech: Covers task design, metrics, and failure cases; useful for model evaluation and regression tests.

Structured Transforms for Low-Overhead Quantization of Language Models: Covers inference cost, latency, throughput, and deployment constraints; useful for systems optimization.

Multimodal Taxonomic Conditioning for Generative Plankton Imagery: Covers retrieval, knowledge-base QA, and evidence reliability; useful as a RAG evaluation lead.

Learnware and AI Model Management System: Covers tool use, execution feedback, and reusable capabilities; useful as an agent reliability lead.

Self-Supervised Cardiac Phase Detection via Single-Parameter Latent Orbits: Covers tool use, execution feedback, and reusable capabilities; useful as an agent reliability lead.

ZipCodec: Ultra-Low-Frame-Rate Streaming Speech Coding: Covers a concrete code intelligence signal; useful as a follow-up candidate.

GPU-CFR: 80x Faster Counterfactual Regret Minimization by Compiling the Game to Static Dataflow and CUDA Graph Replay: Covers a concrete code intelligence signal; useful as a follow-up candidate.

CausalArena: Benchmarking Causal Discovery in the Foundation Model Era: Covers task design, metrics, and failure cases; useful for model evaluation and regression tests.

Nuha-Speech: Building General-Purpose Arabic Speech-LLMs: Covers task design, metrics, and failure cases; useful for model evaluation and regression tests.

CoRA-NAS: Coarse Ranking and Anchor-Residual Refinement for Neural Architecture Search: Covers a concrete training and post-training signal; useful as a follow-up candidate.

On the Regularization Landscape for the Linear Recommendation Models: Covers a concrete training and post-training signal; useful as a follow-up candidate.

Evaluating Time-Series Foundation Models and Multimodal Dietary Context for CGM Forecasting: Covers task design, metrics, and failure cases; useful for model evaluation and regression tests.

Augustinian BabyLM: What Ostensive Definition Can and Cannot Teach a Small Language Model: Covers task design, metrics, and failure cases; useful for model evaluation and regression tests.

Epistemic orientation predicts legislative effectiveness among members of the US Congress: Covers a concrete speech and audio signal; useful as a follow-up candidate.

IndicTriMix: Developing Language Identification Datasets and Models for Tri-Language Code-Mixing: Covers task design, metrics, and failure cases; useful for model evaluation and regression tests.

Thinking with Looped Flows: Covers a concrete training and post-training signal; useful as a follow-up candidate.

Whisper-Based Speech Transcription from Videos Across Multiple Languages for Cross-Cultural Understanding: Covers tool use, execution feedback, and reusable capabilities; useful as an agent reliability lead.

The widening evaluation gap in medical large language model research 2023 to 2026: Covers task design, metrics, and failure cases; useful for model evaluation and regression tests.

Sparsity Regularized and Robust Mean Variance Portfolio Selection Under Ellipsoidal Uncertainty: Covers a concrete code intelligence signal; useful as a follow-up candidate.

ORCH: Organizational Principles Enable Collective Intelligence in Embodied AI: Covers tool use, execution feedback, and reusable capabilities; useful as an agent reliability lead.

Reading boundaries

Automated ranking favors papers with community, code, and applied-engineering signals.
Briefs are based on titles, abstracts, and public metadata by default, not full-paper review.
External API failures degrade optional signals and are reflected in internal records.