archive
Every paper Pith has read. Search by title, abstract, or pith.
1286 papers in cs.IR · page 12
-
Agentic GraphRAG beats vector RAG on Swiss registry queries
Agentic GraphRAG: Navigating Unstructured Financial Data with Collaborative AI
-
LLMs replace rules for BPMN models but accuracy and tests lag
Large Language Models to Enhance Business Process Modeling: Past, Present, and Future Trends
-
Graph-to-text method improves product bundling by up to 26%
Dual-Enhancement Product Bundling: Bridging Interactive Graph and Large Language Model
-
Simulations track AI systems' user retention over repeated interactions
Evaluation of Agents under Simulated AI Marketplace Dynamics
-
Joint profiles lift recommendation accuracy
DUET: Joint Exploration of User Item Profiles in Recommendation System
-
Urgency encodings lift fantasy sports rankings 9% over LightGBM
Driving Engagement in Daily Fantasy Sports with a Scalable and Urgency-Aware Ranking Engine
-
TokenFormer unifies multi-field and sequential recommenders
TokenFormer: Unify the Multi-Field and Sequential Recommendation Worlds
-
RRF fusion leads hybrid retrieval for COVID-19 papers
Hybrid Retrieval for COVID-19 Literature: Comparing Rank Fusion and Projection Fusion with Diversity Reranking
-
Semantic search retrieves HPC tickets despite typos and rewording
FRAGATA: Semantic Retrieval of HPC Support Tickets via Hybrid RAG over 20 Years of Request Tracker History
-
Time-window splits cut data leakage in next-batch rec evaluation
RecNextEval: A Reference Implementation for Temporal Next-Batch Recommendation Evaluation
-
Many-to-many federated training lifts sequential recs in every market
From Transfer to Collaboration: A Federated Framework for Cross-Market Sequential Recommendation
-
Two-stage debate sharpens entity alignment across graphs
Debate to Align: Reliable Entity Alignment through Two-Stage Multi-Agent Debate
-
Authority scoring lifts generative search accuracy and trust
From Relevance to Authority: Authority-aware Generative Retrieval in Web Search Engines
-
Coarse-to-fine time embeddings model actual intervals in user sequences
RoTE: Coarse-to-Fine Multi-Level Rotary Time Embedding for Sequential Recommendation
-
SID alignment cuts retriever training by 8-9 times
Mitigating Collaborative Semantic ID Staleness in Generative Retrieval
-
MLLMs score image pairs via next-token probabilities for zero-shot retrieval
Indexing Multimodal Language Models for Large-scale Image Retrieval
-
LaTeX tool adds citations without leaving the editor
OverCite: Add citations in LaTeX without leaving the editor
-
Content-only encoder ranks cold items more accurately than hybrid methods
Sparse Contrastive Learning for Content-Based Cold Item Recommendation
-
Re-rankers favor old documents over new facts in RAG
FRESCO: Benchmarking and Optimizing Re-rankers for Evolving Semantic Conflict in Retrieval-Augmented Generation
-
Hierarchical index speeds exact retrieval for large rec models
Efficient Retrieval Scaling with Hierarchical Indexing for Large Scale Recommendation
-
AI system nudges tourists to greener destinations with explanations
TRACE: A Conversational Framework for Sustainable Tourism Recommendation with Agentic Counterfactual Explanations
-
Adaptive routing beats fixed retrieval on complex document queries
Adaptive Query Routing: A Tier-Based Framework for Hybrid Retrieval Across Financial, Legal, and Medical Documents
-
Results-only novelty papers out-cite those with all three novelty types
Beyond Single-Dimension Novelty: How Combinations of Theory, Method, and Results-based Novelty Shape Scientific Impact
-
Sliding windows make long-sequence recommenders trainable on modest hardware
Is Sliding Window All You Need? An Open Framework for Long-Sequence Recommendation
-
Knowledge graph RAG lifts accuracy 70% on federal regulations
Knowledge Graph RAG: Agentic Crawling and Graph Construction in Enterprise Documents
-
Situation-aware network improves CTR by modeling user behavior context
Deep Situation-Aware Interaction Network for Click-Through Rate Prediction
-
Attribute prefixes close generative-discriminative gap in recommendation
UniRec: Bridging the Expressive Gap between Generative and Discriminative Recommendation via Chain-of-Attribute
-
Thought retrieval lets LLMs use unlimited external knowledge
Thought-Retriever: Don't Just Retrieve Raw Data, Retrieve Thoughts for Memory-Augmented Agentic Systems
-
Single adversarial document poisons LLM reasoning
AdversarialCoT: Single-Document Retrieval Poisoning for LLM Reasoning
-
Clustering user profiles improves personalized RAG
ClusterRAG: Cluster-Based Collaborative Filtering for Personalized Retrieval-Augmented Generation
-
Agent pipeline creates memory datasets for LLM chat training
AgenticAI-DialogGen: Topic-Guided Conversation Generation for Fine-Tuning and Evaluating Short- and Long-Term Memories of LLMs
-
RAG gains over 25% diversity by indexing opinions explicitly
Retrieval-Augmented Generation Must Move Beyond Factual Grounding to Represent Diverse Opinions
-
Semantic retrieval offers reliable document selection for text analysis
The Effect of Document Selection on Query-focused Text Analysis
-
Parser and chunking choices determine RAG success on financial PDFs
Empirical Evaluation of PDF Parsing and Chunking for Financial Question Answering with RAG
-
Algorithm achieves constant approximation for uniform decision trees
Constant-Factor Approximation for the Uniform Decision Tree
-
RAG system delivers 24/7 citation-backed help for PDB depositors
RCSB PDB AI Help Desk: retrieval-augmented generation for protein structure deposition support
-
Multi-step agent with triple selectors advances entity alignment
EA-Agent: A Structured Multi-Step Reasoning Agent for Entity Alignment
-
Benchmark shows LLMs limited in judging research novelty
NovBench: Evaluating Large Language Models on Academic Paper Novelty Assessment
-
Reference vector stabilizes VAE for semantic item IDs in recommenders
R3-VAE: Reference Vector-Guided Rating Residual Quantization VAE for Generative Recommendation
-
QA trace before generation improves novel character descriptions
Think Before you Write: QA-Guided Reasoning for Character Descriptions in Books
-
Mycelium ANN matches recall using 5x less RAM
Mycelium-Index: A Streaming Approximate Nearest Neighbor Index with Myelial Edge Decay, Traffic-Driven Reinforcement, and Adaptive Living Hierarchy
-
Agent turns plain-English health questions into database queries
ClinQueryAgent: A Conversational Agent for Population Health Management
-
Dual-view reranker selects minimal docs for multi-hop questions at low latency
DualView: Adaptive Local-Global Fusion for Multi-Hop Document Reranking
-
Local LLMs build knowledge graphs zero-shot at 0.70 F1
Frugal Knowledge Graph Construction with Local LLMs: A Zero-Shot Pipeline, Self-Consistency and Wisdom of Artificial Crowds
-
LLM answer snippets clean hard negatives for better retrieval
ARHN: Answer-Centric Relabeling of Hard Negatives with Open-Source LLMs for Dense Retrieval
-
Memory benchmarks cover at most 2 of 7 continuity properties
ATANT v1.1: Positioning Continuity Evaluation Against Memory, Long-Context, and Agentic-Memory Benchmarks
-
Separate spaces for old and new entity knowledge boost link prediction
Multi-Faceted Continual Knowledge Graph Embedding for Semantic-Aware Link Prediction
-
Asymmetric encoders raise Chinese medical retrieval accuracy at unchanged speed
Benchmarking and Enabling Efficient Chinese Medical Retrieval via Asymmetric Encoders
-
Interactive tool turns schema-match validations into live benchmarks
BDIViz in Action: Interactive Curation and Benchmarking for Schema Matching Methods
-
Search tools can close the ethical shopping intention gap
From Query to Conscience: The Importance of Information Retrieval in Empowering Socially Responsible Consumerism