2026-08-23

Make RAG retrieval and knowledge-base QA more reliable, Improve code generation, execution feedback, and automated repair

Today tracks: make RAG retrieval and knowledge-base QA more reliable, make RAG retrieval and knowledge-base QA more reliable, improve code generation, execution feedback, and automated repair.

View sources 中文

Published 2026-08-23 Target source 2026-08-21 Actual source 2026-08-20 Candidates 291 Featured 6 Tracked 20 Source date fallback

This issue fetched and deduplicated 291 candidate papers from the 2026-08-20 source date, then selected 6 featured papers and 20 additional mentions.

Featured

1Manifold Drift in Flow Preference Optimization: A Root Cause of Reward Hacking🔗
2Scale-Aware Pretraining of Time Series Foundation Models via Multi-Patch Token Alignment and Hybrid Masking🔗
3Open-Vocabulary 3D Object Detection with Co-Distillation Discovery and Dual Guidance Robust Training🔗
4Multi-Source Wasserstein Distributionally Robust Graph Learning🔗
5A knowledge-guided agentic framework for mitigating patient-context ambiguity in health queries🔗
6SWE-bench Science: Can Coding Agents Resolve Engineering Tasks in Science?🔗

What is worth tracking today

Today’s high-signal papers point to: make RAG retrieval and knowledge-base QA more reliable, make RAG retrieval and knowledge-base QA more reliable, improve code generation, execution feedback, and automated repair. 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

make RAG retrieval and knowledge-base QA more reliable

Manifold Drift in Flow Preference Optimization: A Root Cause of Reward Hacking (Yansen Han, Shengyi Liao, Yuanxing Zhang, Pengfei Wan, Tao Lin) 2608.20011 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 rag, alignment, benchmark, fine-tuning to frame the training and post-training 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: Preference optimization is a standard alignment method for generative models, yet extending it to continuous-time dynamics remains non-trivial

Keywordsragalignmentbenchmarkfine-tuning

Code/DataCheck the source paper

make RAG retrieval and knowledge-base QA more reliable

Scale-Aware Pretraining of Time Series Foundation Models via Multi-Patch Token Alignment and Hybrid Masking (Taihua Chen, Xiang Ma, Yixin Zhang, Tailin Zhan, Manyu Sun, Lizhen Cui) 2608.20005 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 rag, serving, alignment, benchmark to frame the benchmarks and evaluation 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: Pretraining time series foundation models across heterogeneous datasets necessitates effective handling of varying sampling frequencies

Keywordsragservingalignmentbenchmark

Code/DataCheck the source paper

improve code generation, execution feedback, and automated repair

Open-Vocabulary 3D Object Detection with Co-Distillation Discovery and Dual Guidance Robust Training (Shangbo Yuan, Jie Xu, Xiaofeng Zhu, Na Zhao) 2608.19973 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 alignment, code, training, distillation 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: Recently, open-vocabulary 3D object detection (3D-OVD) has gained increasing attention for its ability to detect unseen objects in 3D scenes

Keywordsalignmentcodetrainingdistillation

Code/DataCheck the source paper

make RAG retrieval and knowledge-base QA more reliable

Multi-Source Wasserstein Distributionally Robust Graph Learning (Chuansen Peng, Yifan Xia, Jinshan Zhong, Xiaojing Shen) 2608.19914 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 rag, inference, serving, benchmark to frame the benchmarks and evaluation 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: Network topology inference from graph signals is central to graph signal processing with applications in neuroscience, sensor, and social networks

Keywordsraginferenceservingbenchmark

Code/DataCheck the source paper

make agents use tools and reusable skills more reliably

A knowledge-guided agentic framework for mitigating patient-context ambiguity in health queries (Mahyar Abbasian, Saba A. Farahani, Arshia Ilaty, Hung Cao, Ramesh Jain, Amir M. Rahmani) 2608.19875 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, retrieval, safety, benchmark to frame the retrieval and rag 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: Patients often submit short, underspecified queries to healthcare chatbots that lack the patient-specific information needed to determine an appropriate response

Keywordsagentretrievalsafetybenchmark

Code/DataCheck the source paper

make agents use tools and reusable skills more reliably

SWE-bench Science: Can Coding Agents Resolve Engineering Tasks in Science? (Zhipeng Xu, Jiahao Lu, Yining Zheng, Yuxin Wang, Xipeng Qiu) 2608.19799 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, serving, evaluation to frame the code intelligence 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: Software increasingly functions as part of the scientific instrument itself, making failures in scientific code capable of compromising not only program behavior but also the evidence underlying scientific conclusions

Keywordsagentragservingevaluation

Code/DataCheck the source paper

Other papers worth tracking

DIFFCZSL: Compositional Zero-Shot Learning Regularized by Diffusion Representations: Covers retrieval, knowledge-base QA, and evidence reliability; useful as a RAG evaluation lead.

Towards Clinically Faithful Medical Image Captioning via Enhanced Vision-Language Alignment: Covers model safety, guardrail routing, risk classification, or governance evaluation; useful as a safety workflow lead.

LoRA-GA$^2$: Low Rank Adaptation with Multi-step Gradient Adaptive Alignment: Covers model safety, guardrail routing, risk classification, or governance evaluation; useful as a safety workflow lead.

RecPFN: Prior-Fitted Networks for In-Context-Based Recommendations: Covers task design, metrics, and failure cases; useful for model evaluation and regression tests.

SafeBranch: Branch-Pair Safety Alignment for Embodied Agents: Covers model safety, guardrail routing, risk classification, or governance evaluation; useful as a safety workflow lead.

Scale-Separated Conditioning for Style-Encoder-Free Diffusion Stylization: Covers a concrete vision and image generation signal; useful as a follow-up candidate.

Robust Cross-Modal Foundation Model Perception for Underwater Robots under Degraded Visual Conditions: Covers a concrete multimodal models signal; useful as a follow-up candidate.

Frequency-Aware Continual Learning for Smart Contract Vulnerability Detection with Large Language Models: Covers retrieval, knowledge-base QA, and evidence reliability; useful as a RAG evaluation lead.

ReCache: Efficient KV Cache Reuse and Compression for Tool-Augmented LLM Agents: Covers tool use, execution feedback, and reusable capabilities; useful as an agent reliability lead.

Scientific Data Skills: Enabling Agent-Ready Scientific Data Services at Scale: Covers retrieval, knowledge-base QA, and evidence reliability; useful as a RAG evaluation lead.

Mitigating Identity Essentialism in LLM Agents with Longitudinal Life Trajectories: Covers tool use, execution feedback, and reusable capabilities; useful as an agent reliability lead.

PEA-DPO: Perception-Enhanced Alignment Direct Preference Optimization for MLLMs Alignment: Covers model safety, guardrail routing, risk classification, or governance evaluation; useful as a safety workflow lead.

OrthoSkillVLA: Continual Skill Learning via Gradient-Informed Skill Subspace Adaptation: Covers a concrete training and post-training signal; useful as a follow-up candidate.

Enforcing LLM Safety through DMD-based Classification of Prompt-Response Embedding Dynamics: Covers model safety, guardrail routing, risk classification, or governance evaluation; useful as a safety workflow lead.

CVSD-Reg: Cross-Modal Visual Semantic Prior Distillation for Robust LiDAR Registration: Covers a concrete reasoning and planning signal; useful as a follow-up candidate.

Automated Summarization of Financial News Using Large Language Models and Retrieval-Augmented Generation: An Early Empirical Study (Fall 2023): Covers retrieval, knowledge-base QA, and evidence reliability; useful as a RAG evaluation lead.

A Standardized Framework for Machine Learning in Power System Protection: Covers retrieval, knowledge-base QA, and evidence reliability; useful as a RAG evaluation lead.

Task-CoEvolve: Efficient Harness Optimization via Adaptive Validation Task Selection: Covers tool use, execution feedback, and reusable capabilities; useful as an agent reliability lead.

ExPhy: A Benchmark for Explicit Physical Property Learning in Multi-Object Trajectory Forecasting: Covers tool use, execution feedback, and reusable capabilities; useful as an agent reliability lead.

HealMed: Multilingual Evaluation of Large Language Models in Medicine: Covers task design, metrics, and failure cases; useful for model evaluation and regression tests.

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.