archive
Every paper Pith has read. Search by title, abstract, or pith.
738 papers in cs.IR · page 1
-
Citations miss key context in agent graph answers
Why Neighborhoods Matter: Traversal Context and Provenance in Agentic GraphRAG
-
Optimal logging policies minimize OPE error via reward-coverage balance
Logging Policy Design for Off-Policy Evaluation
-
The paper presents a fixed six-stage deterministic workflow that confines language model…
A Deterministic Agentic Workflow for HS Tariff Classification: Multi-Dimensional Rule Reasoning with Interpretable Decisions
-
Aggregated vectors make different financial docs look identical
A Picture is Worth a Thousand Words? An Empirical Study of Aggregation Strategies for Visual Financial Document Retrieval
-
AsymRec raises generative recommender accuracy 15.8%
Asymmetric Generative Recommendation via Multi-Expert Projection and Multi-Faceted Hierarchical Quantization
-
Distilled rerankers match quality with 34% fewer tokens
Stop Overthinking: Unlocking Efficient Listwise Reranking with Minimal Reasoning
-
Adaptive gate skips reasoning for simple multimodal inputs
Think When Needed: Adaptive Reasoning-Driven Multimodal Embeddings with a Dual-LoRA Architecture
-
Semantic IDs halve beam search size for e-commerce retrieval
Efficient Generative Retrieval for E-commerce Search with Semantic Cluster IDs and Expert-Guided RL
-
PaSaMaster beats GPT-5.2 in paper retrieval at 1% cost
Towards Self-Evolving Agentic Literature Retrieval
-
Imagined future steps triple recall of distant memories
Thinking Ahead: Prospection-Guided Retrieval of Memory with Language Models
-
Small rotations hide data in embeddings undetected
VectorSmuggle: Steganographic Exfiltration in Embedding Stores and a Cryptographic Provenance Defense
-
The paper describes benchmarks of XRootD and Pelican services in the Open Science Data…
Benchmarking the Open Science Data Federation services to develop XRootD best practices
-
Granite R2 models lead multilingual retrieval in 200+ languages
Granite Embedding Multilingual R2 Models
-
LLM profiles boost recommender simulation ranking by 7%
Task-Aware Automated User Profile Generation for Recommendation Simulation Using Large Language Models
-
Graph links convergent claims from multiple innovation methods
IdeaForge: A Knowledge Graph-Grounded Multi-Agent Framework for Cross-Methodology Innovation Analysis and Patent Claim Generation
-
Graph links 200k research repos to papers and artifacts
SemRepo: A Knowledge Graph for Research Software and Its Scholarly Ecosystem
-
Parallel dataset gives medical dialogues in nine Indic languages
IndicMedDialog: A Parallel Multi-Turn Medical Dialogue Dataset for Accessible Healthcare in Indic Languages
-
Latent info gain ranks visual evidence for better multimodal RAG
Utility-Oriented Visual Evidence Selection for Multimodal Retrieval-Augmented Generation
-
LeanSearch v2 lifts Lean 4 proof success to 20 percent
LeanSearch v2: Global Premise Retrieval for Lean 4 Theorem Proving
-
LeanSearch v2 lifts Lean 4 proof success to 20%
LeanSearch v2: Global Premise Retrieval for Lean 4 Theorem Proving
-
Multi-agent system automates VC due diligence
A Multi-Agent Orchestration Framework for Venture Capital Due Diligence
-
Half of ReDial CRS accuracy traces to repetition shortcuts
A Standardized Re-evaluation of Conversational Recommender Systems on the ReDial Dataset
-
LLMs predict query-specific validity horizons for web content
RAG-Enhanced Large Language Models for Dynamic Content Expiration Prediction in Web Search
-
Source figures become verifiable evidence in deep research reports
ViDR: Grounding Multimodal Deep Research Reports in Source Visual Evidence
-
KITE tutor raises simulated student accuracy on algorithm tasks
Retrieval-Augmented Tutoring for Algorithm Tracing and Problem-Solving in AI Education
-
Context changes what the same image means for retrieval
Same Image, Different Meanings: Toward Retrieval of Context-Dependent Meanings
-
Linked page ecosystems steer LLM agents to target recommendations
EcoGEO: Trajectory-Aware Evidence Ecosystems for Web-Enabled LLM Search Agents
-
MLP distillation accelerates generative recommenders 8.74 times
MLPs are Efficient Distilled Generative Recommenders
-
Admins like AI help writing WhatsApp rules but fear trust breaches
Creating Group Rules with AI: Human-AI Collaboration in WhatsApp Moderation
-
LLM refines embeddings at test time for up to 25% gains
Task-Adaptive Embedding Refinement via Test-time LLM Guidance
-
This paper proposes ORBIT, a method that tracks how far a fine-tuned generative retrieval…
ORBIT: Preserving Foundational Language Capabilities in GenRetrieval via Origin-Regulated Merging
-
Entropy of plausibility scores estimates LLM question difficulty
Question Difficulty Estimation for Large Language Models via Answer Plausibility Scoring
-
High-convergence sentences lift LLM accuracy on inferential questions
Context Convergence Improves Answering Inferential Questions
-
Benchmark forces models to combine facts from two articles
MedHopQA: A Disease-Centered Multi-Hop Reasoning Benchmark and Evaluation Framework for LLM-Based Biomedical Question Answering
-
Prototype-guided retrieval improves EHR clinical predictions
EHR-RAGp: Retrieval-Augmented Prototype-Guided Foundation Model for Electronic Health Records
-
Retrieval lifts two-hop medical QA to 89% conceptual accuracy
Overview of the MedHopQA track at BioCreative IX: track description, participation and evaluation of systems for multi-hop medical question answering
-
BatchBench framework equalizes autoscaling policy tests
BatchBench: Toward a Workload-Aware Benchmark for Autoscaling Policies in Big Data Batch Processing -- A Proposed Framework
-
Crowdsourcing validates LLM ontology mappings at scale
Unlocking Crowdsourcing for Ontology Matching Validation
-
One autoregressive model makes personalized ad images and text
Design Your Ad: Personalized Advertising Image and Text Generation with Unified Autoregressive Models
-
Three-stage retrieval pipeline ranks 8th in SemEval multi-turn task
Caraman at SemEval-2026 Task 8: Three-Stage Multi-Turn Retrieval with Query Rewriting, Hybrid Search, and Cross-Encoder Reranking
-
Health record trajectories improve image-based disease forecasts
From Trajectories to Phenotypes: Disease Progression as Structural Priors for Multi-organ Imaging Representation Learning
-
Ulam similarity admits O(n/sqrt(log n)) LSH distortion
On the LSH Distortion of Ulam and Cayley Similarities
-
Benchmark with 1M entries tests multi-dimensional rewards for recommender agents
RecRM-Bench: Benchmarking Multidimensional Reward Modeling for Agentic Recommender Systems
-
ZipRerank matches top multimodal rerankers at 10x lower latency
Very Efficient Listwise Multimodal Reranking for Long Documents
-
Critic and generator agents iteratively refine research outlines
AgentDisCo: Towards Disentanglement and Collaboration in Open-ended Deep Research Agents
-
Dual-context views with quality weights boost sequential recs
Quality-Aware Collaborative Multi-Positive Contrastive Learning for Sequential Recommendation
-
Staged mining and activity grouping boost LLM recommendations
HSUGA: LLM-Enhanced Recommendation with Hierarchical Semantic Understanding and Group-Aware Alignment
-
Planner picks slow reasoning only when it improves recommendations
TwiSTAR:Think Fast, Think Slow, Then Act,Generative Recommendation with Adaptive Reasoning
-
Conditional memory fixes SID representation conflicts in generative recommendation
Conditional Memory Enhanced Item Representation for Generative Recommendation
-
Codebooks quantize signals to boost multi-market CTR privately
FedMM: Federated Collaborative Signal Quantization for Multi-Market CTR Prediction