A team called VANGUARD describes IROH, a three-stage retrieval pipeline that blends sparse and dense retrieval with cross-encoder reranking and LLM judges distilled from rationales. The system placed first on the English JOKER 2026 Task 1 leaderboard with a MAP score of 0.6347. The work appears as an arXiv paper under the cs.CL category.
A new arXiv paper addresses how large language models tend to ignore information placed in the middle of long contexts, a problem studied mostly for retrieval tasks rather than long-input-to-long-output generation. The authors introduce a synthetic dataset and evaluation framework for this setting and propose a mitigation approach. The work is a revised cross-listing (v2) on arXiv cs.AI.
A new arXiv preprint argues that the two most common query-document scoring methods, TF-IDF and BM25, both correspond exactly to Kullback-Leibler divergences. This gives the widely used retrieval heuristics a probabilistic justification and places them inside a single statistical framework. The author frames the work as filling a long-standing gap in how these ranking functions are derived.
A new arXiv preprint examines whether decoder-only language models built with mixture-of-experts layers can serve as effective first-stage retrievers. The authors note that retrieval quality typically rises with backbone size, but dense models force every query and document through all parameters. The work targets strong retrieval performance while limiting the computational cost per query.
A new arXiv paper introduces CompCQR, a method that rewrites ambiguous, context-dependent user utterances into clearer retrieval queries for multi-turn conversational search. The approach composes queries without requiring task-specific training, aiming to bridge the gap between conversational phrasing and standard retrieval systems. The work targets information-seeking dialogue, where follow-up questions often lack the context needed for direct use as search queries.
A revised arXiv paper proposes sublinear sketching techniques for two core machine learning problems: approximate nearest neighbor search and approximate kernel density estimation. The approach targets large-scale data analysis and information retrieval settings where exact computation is impractical, aiming to reduce memory and query costs while preserving accuracy guarantees.
A revised arXiv paper addresses a known weakness in dense retrieval: compressing a passage into a single embedding tends to weight early text more heavily, so retrieval quality drops when the relevant span sits later in the passage. The authors propose attention calibration as a way to reduce this positional bias, building on earlier inference-time approaches. It is a research contribution rather than a released product.
A new arXiv paper examines how news articles often reference dates incompletely, such as mentioning only a day or month, which requires knowledge of the publication date to interpret correctly. This becomes problematic when articles are indexed as standalone text in search and retrieval-augmented applications. The work, focused on Japanese news, proposes methods for reconstructing these omitted temporal expressions.
A new study tests whether rewriting a user's question into multiple search variants actually improves retrieval-augmented generation when the retrieval system is already highly capable. The authors compare several rewriting strategies within a single fixed evaluation setup to measure what each adds over a strong baseline.
NeoMME is a newly introduced encoder model designed to natively process multiple input modalities while supporting a broad range of languages. Its architecture emphasizes efficiency, aiming to deliver competitive performance at lower computational cost for workloads such as embeddings and retrieval. The release is detailed on the Hugging Face blog, making the model accessible to developers building multimodal, multilingual applications.
WHY IT MATTERS ↘For teams running embedding and retrieval pipelines, a single efficient encoder covering many languages and modalities could reduce inference costs and simplify architectures that otherwise chain separate per-modality or per-language models. It also signals growing competition in a segment long dominated by English-centric encoders, potentially lowering the barrier to shipping multilingual, multimodal search applications.