TALRanker is a novel framework that formalizes pointwise relevance scoring as an agentic Markov decision process that achieves state-of-the-art performance across standard and reasoning-intensive retrieval benchmarks, matching throughput with pointwise rerankers while outperforming parameter-heavy reasoning models.
Abstract
Generative Large Language Models (LLMs) have revolutionized information retrieval, yet their strictly parametric nature frequently leads to severe factual hallucinations when confronted with complex queries beyond their epistemic boundaries. While external tool-calling can mitigate this, indiscriminately invoking search tools for every document during reranking incurs prohibitive latency overheads, creating an intractable accuracy-efficiency dilemma. To address this challenge, we propose TALRanker, a novel framework that formalizes pointwise relevance scoring as an agentic Markov decision process. We optimize it via a two-stage training paradigm. An initial warm-up utilizes a language-preserving hybrid loss to prevent the catastrophic forgetting of native generative capacities. Subsequently, an asymmetric cost-aware reward equipped in reinforcement learning forces the policy to autonomously bypass tools for maximum efficiency when confident, while selectively retrieving external evidence to avert severe hallucination penalties when uncertain. Extensive evaluations demonstrate that TALRanker achieves state-of-the-art performance across standard and reasoning-intensive retrieval benchmarks, matching throughput with pointwise rerankers while outperforming parameter-heavy reasoning models.
Sparsely-activated Mixture-of-Experts (MoE) language models achieve remarkable inference efficiency by activating only a small fraction of parameters per token, yet their full expert banks reside in memory at all times, creating a prohibitive deployment bottleneck. Existing structured pruning methods, largely designed for dense transformers, assess expert importance using locally derived heuristics that are blind to the interdependent nature of MoE routing. We introduce MAESTRO (Markov-chain Approximated Expert Sparsification via Transition-based ROuting), a structured pruning framework designed for MoE architectures that models autoregressive expert activation trajectories as Ergodic Markov chains whose stationary distributions encode cross-layer dependencies, yielding a globally aware importance heuristic. Evaluated across five diverse domains including Safety, Bias, and Ethics, MAESTRO outperforms state-of-the-art baselines by up to 10.61% in average performance retention under a strict 50% compression regime, while exhibiting substantially lower cross-task variance, indicating that global, routing-congruent pruning produces models that generalize more consistently across heterogeneous tasks.
Multimodal Document Question Answering (MDQA) systems commonly follow a retrieve-then-answer paradigm; however, the retrieval stage often introduces substantial noise, making an effective reranking component indispensable. Existing reranker training frameworks in MDQA rely predominantly on proxy supervision derived from human annotations or large language model (LLM) outputs, which are frequently noisy and, more critically, misaligned with downstream answer quality. To overcome this limitation, we propose AD-Reranker, a novel framework that shifts reranker training from proxy imitation to answer-driven utility optimization. Specifically, we reformulate the reranker as an environment-grounded agent that interacts with a downstream reader, modeled as a deterministic environment. We further design a composite reward function that integrates answer correctness, thereby explicitly incentivizing ranking strategies aligned with downstream task performance. To optimize the agent, we adopt Group Relative Policy Optimization (GRPO), enabling stable and effective group-wise policy learning. Empirical results demonstrate that AD-Reranker achieves superior reranking quality and an optimal accuracy-efficiency trade-off. When integrated into standard MDQA pipelines, AD-Reranker improves accuracy by 1.9%–5.0% while reducing the reader's context token consumption by 15%–52%, providing strong evidence for the effectiveness of answer-driven reranker training.
Keyu Zhu, Shuanghong Shen, Xianquan Wang et al.· Annual International ACM SIG...· 0 citations
The rapid adoption of generative search engines has marked a significant shift in information retrieval. New approaches leverage Large Language Models (LLMs) to provide synthesized, contextually rich responses in natural language to directly satisfy users’ information needs. Due to their generative nature, these systems benefit from enhanced reliability by integrating explicit references to supporting sources in their outputs, otherwise known as Attribution. Teaching LLMs this capability is a challenging task that heavily depends on training data quality and strong training signals, especially for scenarios that require digesting long contexts and providing comprehensive answers. Planning and decomposition have notoriously helped language models produce better answers, but their impact on attribution remains relatively unexplored. This work proposes RAGnRoll, a language model for attributed answer generation within a multi-round Retrieval-Augmented Generation (RAG) framework. We leverage the iterative nature of multi-round RAG to train an LLM to incrementally build answers guided by subqueries. RAGnRoll is trained using a causal language modeling objective and a small, specially crafted dataset. Empirical evaluations on four different datasets for attributed long answer generation (ASQA, ELI5) and information-seeking (HAGRID and CORAL) demonstrate RAGnRoll's impact on both answer correctness and attribution accuracy. We extensively evaluate RAGnRoll training approach against baseline methods and benchmark its performance at inference against strong state-of-the-art models. RAGnRoll achieves competitive performance with advanced approaches without relying on extensive data or complex training.
Hanane Djeddal, Laure Soulier, K. Pinel-Sauvagnat et al.· ACM Transactions on Informat...· 0 citations
Large Language Models (LLMs) have emerged as powerful assets for recommender systems. However, deploying them as generative recommenders or zero-shot rankers at web-scale remains bottlenecked by prohibitive computational overhead and grounding challenges. In this paper, we revitalize the classic, highly efficient two-tower retrieval architecture by adapting LLMs as semantic representation backbones rather than generative engines. We introduce an LLM-native two-tower framework engineered for high-throughput, large-scale retrieval. Our architecture introduces several key innovations: a shared LLM encoder for joint user-item modeling, End-Of-Sentence (EOS) token pooling for compact sequence embedding, cross-dataset transfer learning, knowledge distillation from powerful cross-encoder teachers, and latent reasoning within the user tower. Extensive evaluation across three public benchmarks demonstrates that cross-encoder architecture outperforms current state-of-the-art (SoTA) models, while the efficient two-tower student achieves SoTA-comparable retrieval performance. Furthermore, experiments on internal large-scale production systems yield substantial topline retrieval improvements along with high resilience to model staleness and superior data scaling. Our findings demonstrate that when augmented with modern representation learning, the traditional two-tower paradigm remains an exceptionally competitive and practical solution for industrial retrieval systems.
Zhe Xu, Prachi Agrawal, Kavosh Asadi et al.· 0 citations
Large language models (LLMs) are increasingly deployed in high-stakes settings, yet reliably estimating when their outputs should be trusted remains an open challenge. Existing uncertainty estimation approaches—such as calibration, token-level probabilities, or semantic entropy—typically require access to model internals, additional supervision, or computationally intensive pipelines. We propose answer instability, defined as the variability of a model’s final answer across repeated stochastic generations of the same prompt, as a simple, label-free, and black-box uncertainty signal. Evaluated across three task types — reasoning, multiple-choice QA, and constraint-following — using four LLMs and 520 prompt-model pairs, our approach achieves performance competitive with semantic entropy while requiring no semantic similarity model. Our results show that instability strongly correlates with prediction errors and reliably discriminates correct from incorrect outputs. We further demonstrate its utility for selective prediction and targeted repair, improving reliability without access to internal probabilities or additional training.
J. M. Araña, Kristine Ann M. Carandang, E. R. Casin et al.· Annual Meeting of the Associ...· 0 citations
It is found that, in this benchmark, supervised fine-tuning (SFT) provides a strong baseline, substantially improving argument language consistency and end-to-end function call accuracy and, under consistent model selection, SFT achieves performance comparable to, and sometimes exceeding more complex reinforcement learning (RL) approaches.
Siddharth Chauhan, Thomas Butler, Abhishek Singhania et al.· 0 citations