Negar Arabzadeh

FLINT: Fixed-budget Late Interaction

This work introduces FLINT (Fixed-budget Late INTeraction), an architecture that uses attention-based pooling to represent each query and document with a fixed number of embedding vectors, and demonstrates that FLINT's effectiveness-efficiency benefits extend across encoder architectures.

Patient Bytes: A Reliability-Budgeted Scheduler for Agentic LLM Workflows

Seek: Self-Evaluative Exploration for Knowledge Retrieval

seek, Self-Evaluative Exploration for Knowledge Retrieval, a training-free framework that addresses this limitation through iterative corpus interaction at test time through iterative corpus interaction at test time.

Route Me If You Can: A Benchmark for Query Reformulation Selection

QueryRoute is introduced, a benchmark that freezes the expensive artifacts needed to study this inference-time decision problem reproducibly: original queries, generated variants, ranked lists under multiple retrievers, retrieval scores, and per-query oracle labels.

Peerify: Benchmarking Peer-Review Claim Verification

The results demonstrate the importance of retrieval-centered verification and claim decomposition, while highlighting the challenges posed by ambiguous and interpretive reviewer claims.

Judging a Review by its Cover: A Reliability Analysis of LLM-based Peer Review Evaluation Metrics

Findings show that many peer-review evaluation metrics partially conflate review quality with linguistic presentation, and indicate that robustness to meaning-preserving rewriting should be validated before such metrics are used to compare human-written, AI-assisted, and AI-generated reviews.

Question's Gambit: The First Move Matters in Agentic Deep Search

Question's Gambit is introduced, a first-move retrieval module that decomposes the question into a set of clues, reformulates them into complementary searches, consolidates the retrieved results, and reranks the candidate pool before the agent begins its iterative search-and-reasoning process.

RAG over Thinking Traces Can Improve Reasoning Tasks

The results suggest that thinking traces are an effective retrieval corpus for reasoning tasks, and transforming them into structured, compact, or diagnostic representations unlocks even stronger gains.

From Doxa to Logos in Scientific Peer Review

Reviewerly's retrieval-centered infrastructure for auditing peer review at scale is presented and three deployed systems are described, including three deployed systems that analyzes hybrid human--AI authorship by disentangling the origin of ideas from the origin of text in peer reviews, enabling measurement of how hum...