Papers

Recorded lecture videos, often enhanced with search and summarization features, are a standard study resource. However, students cannot easily ask course specific questions or verify answers against an instructor's lecture. We report a semester-long deployment of VideoPoints platform with a retrieval-augmented chatbot that answers from course lecture materials and returns timestamped citations. The chatbot retrieves only from the active course, uses chapter summaries to guide transcript ranking, and returns clickable timestamped citations. Students used it for quick lookups and exam review. Across 833 messages, 70.5% included citations, none crossed a course boundary, and when no lecture evidence matched, the chatbot usually declined rather than answering. Among the users, citations were the most consistently useful feature, while practice-question generation was the strongest unmet request. We also evaluated the design on the real-world test split of EduVidQA, a public multimodal benchmark for lecture-video question answering. Our design improved correct-lecture retrieval by 6.3 percentage points over dense-only retrieval. Together, the results show that effective deployment depends on course isolation, supported citations, and alignment with students' study practices.

Dynamic maximum inner-product search (MIPS) returns the $K$ stored vectors with the largest dot products with a query while allowing the dataset to change through insertions, replacements, and deletions. For edge retrieval, the challenge is to achieve high recall and fast queries without making updates expensive. Full-vector scanning keeps updates simple but compares each query with every stored vector, while indexed methods reduce query cost at the expense of maintaining additional structures during updates. We propose ETAR, an index-free method that reduces query work while preserving simple updates. ETAR keeps the query coordinates with the largest squared values until they cover most of its total squared magnitude and treats the rest as a low-magnitude tail. It estimates similarity from the retained coordinates using a compact lower-precision representation, corrects for skipped coordinates, and reranks a fixed number of candidates using full-precision vectors. Across five runs on nine static datasets, ETAR averages 99.2% Recall@10, the fraction of exact top-10 results recovered, while running over 4$\times$ faster than exact scanning at a representative setting. This speedup also extends to an ARM-based mobile device, where ETAR is up to 6.9$\times$ faster across four synthetic distributions. Under five streaming workloads, it maintains 100% Recall@10 at every measured point without index rebuilds. Overall, ETAR offers a practical middle ground for dynamic MIPS by reducing query cost while retaining simple, index-free updates. Code is available at https://github.com/arasyi/etar-mips.

Large language models (LLMs) achieve state-of-the-art generative ranking quality, but the ranking they produce must be decoded, and autoregressive decoding spends one sequential forward pass per emitted token. We observe that the only tokens a ranker must emit are the $N$ ordinal values naming the items in ranked order, and that this narrow, permutation-structured output format admits decoding strategies which are much more efficient than left-to-right generation. We introduce SPD (Single Forward Pass), a format-specialized decoding strategy that decodes all $N$ ordinals in $O(1)$ forward passes. SPD reads an $N \times K$ item-position score matrix off the LLM's prefill hidden states with a lightweight self-attention head, then decodes the ordinals as the optimal bipartite assignment of that matrix via the Hungarian algorithm, yielding a valid permutation by construction rather than by repair. Through a systematic study of training signals and backbone adaptation, we show that LoRA-based fine-tuning combined with auto-regressive LLM ranking distillation reaches 28 ms end-to-end inference, a speed-up of 64x while maintaining ranking quality on par with the teacher. We provide a complete ablation decomposing the contributions of architecture, training signal, and backbone adaptation. Our framework connects generative ranking to combinatorial optimization, opening a path toward other $O(1)$-decode mechanisms for real-time ranking.

Knowledge Graph Question Answering (KGQA) over RDF graphs remains challenging in domain-specific settings, where formal ontologies and curated text-SPARQL pairs are often unavailable. We present KGVoyager, a KG-agnostic agentic architecture that generates SPARQL queries from natural language questions by dynamically discovering graph structure and semantics, requiring only a query endpoint of the underlying graph. Using a think-act-observe loop with search, exploration, and execution tools, KGVoyager maps terms to graph IRIs, uncovers structure, and refines queries through execution feedback - all without pre-existing ontologies or examples. Unlike the prior state of the art, KGVoyager requires only a lightweight class index which renders it applicable for far more real-world endpoints. Across four benchmarks, KGVoyager improves F1 by ~8 points while cutting cost and runtime by ~22% each.

We evaluate embedding retrieval where surface form and meaning are pulled apart on purpose: retrieving items that share underlying structure but not wording, in two unrelated domains under one protocol, competition mathematics (MathNet-Retrieve; 500 queries, 117,088-item corpus) and embodied-agent trajectories (ALFWorld-derived; 118 queries, 336 trajectories). In mathematics the failure is complete: strict Hit@1 at the heaviest disguise tier is 0.0% for both production embedders (bootstrap 95% CI [0.0, 0.0]) while the correct item sits in the top 10 nearly always, and in 95.2 to 99.8% of misses the winner is more lexically similar to the query than the correct answer. In trajectories, where surface variation is incidental, the same models land at or near hypergeometric chance when gold must involve a different object, and below chance for all three embedders once gold must differ in object and receptacle: retrieval anchors on literal tokens, not task structure. A lexical reranker control hurts in mathematics and helps in trajectories (closing 26 to 36% of the gap, CIs excluding zero); its sign reveals whether a benchmark's surface variation is adversarial or incidental. An LLM reranker recovers 5 to 63% of the gap in mathematics and 43 to 76% in trajectories; direction replicates across three judges (all 21 cells positive), but effect sizes, tier profiles, and the outlier judge change with domain (paired differences excluding zero everywhere). Mathematics gains concentrate on well-known competitions (+19.8 points, CI [+6.7, +33.2], one of six cells), so part of the recovery is memorization. In a paired downstream experiment (210 queries, graders at 96 to 99% agreement), oracle retrieval was indistinguishable from adversarially bad retrieval (McNemar p = 0.678); the solver's 69.5% zero-shot accuracy is largely a truncation proxy (97 to 100% on finished answers), leaving no headroom.

Composed image retrieval (CIR) retrieves a target image from a reference image and a text modification. This paper studies metadata-available CIR reranking, where a fixed CIR model first returns a candidate pool and gallery metadata is then used for second-stage concept-guided scoring. We introduce AutoConcept, a training-free reranker that converts concept evidence into an interpretable memory. AutoConcept filters noisy concepts, activates query-relevant positive constraints with an auxiliary negative penalty, and combines base retrieval scores with metadata-based concept-candidate alignment through inference-time calibration. On FashionIQ, AutoConcept yields significant early-rank improvements over WeiMoCIR and consistent plug-in gains on LinCIR candidate pools. Metadata-aware controls show that structured concept memory adds signal beyond direct query-text and extracted-attribute matching, while a query-only variant further supports the effectiveness of concept-level reranking. A supplementary real-human concept-label study indicates that the same memory interface can consume participant-provided evidence. These results position AutoConcept as an interpretable concept-memory reranker for product-style CIR galleries with available metadata.

Neural ranking models have become core components of modern information retrieval systems and important building blocks of AI systems such as retrieval-augmented generation (RAG) pipelines. However, their robustness remains insufficiently understood in the presence of large language models (LLMs), which can generate fluent and deceptive content at scale. This work investigates the vulnerability of neural ranking models to corpus poisoning attacks, in which an adversary injects a small number of maliciously crafted documents into the corpus to distort ranking behavior. We propose VerTox, the first framework to formulate corpus poisoning as a verifiable reward-guided reinforcement learning (RLVR) problem. By explicitly coupling ranking distortion with factual corruption through specialized reward shaping, we fine-tune compact LLMs into adversarial generators. Experiments demonstrate that our method achieves near-perfect attack success rates, producing adversarial documents that frequently rank higher than target documents across major neural ranking architectures, as well as a proprietary commercial embedding model. The generated adversarial documents are fluent and exhibit low perplexity, making them difficult to detect. Furthermore, by explicitly encouraging factual corruption, our adversarial documents significantly degrade the performance of a downstream RAG application.

Retrieval over visually rich documents has a representation problem: important content often lives in tables, charts, figures, and layout relations that plain OCR linearizes, corrupts, or omits. ColPali-family visual retrievers address this with patch-level multi-vector indexes and late-interaction scoring, keeping image-derived retrieval on the query-time serving path. We introduce MIDR (Multimodal Indexing for Document Retrieval), a training-free framework for enrichment-augmented indexing that shifts multimodal reasoning to index time. During ingestion, a multimodal LLM converts rendered pages into verified textual fields that are indexed with BM25F and optionally fused with dense retrieval, enabling text-centric serving over multimodally grounded evidence. On ViDoRe V3, MIDR Hybrid achieves 0.6219 average nDCG across five English domains, a 23.0% relative gain over BM25, remaining competitive with ColQwen2.5. On two French-document domains, enrichment bridges English queries and French page text, lifting BM25 from 0.1532 to 0.5448 nDCG and outperforming ColQwen2.5. Across all seven domains, MIDR leads ColQwen2.5 on four while using approximately 9x smaller index memory and approximately 2x lower query latency. These results establish index-time multimodal reasoning as a compelling accuracy-deployment alternative to serving-time visual late interaction.

Scaling Transformers has driven large gains in language modeling, but transplanting this to behavior-sequence modeling in production ranking is challenging: recommendation differs in signal quality, where behavior sequences are noisy, temporally irregular, and sparsely supervised, and in computation asymmetry, where each request scores many candidates against one shared user history under tight latency budgets. We propose ReST, a recommendation-native Transformer scaling framework. For signal quality, it introduces a sequence encoder with dual-gated attention, rotary positional and temporal embedding, stabilized residual normalization, and training-only auxiliary objectives. For computation asymmetry, it factorizes ranking into a heavy reusable encoder and a lightweight cross decoder with projection-free KV attention and token-specific parameterization, coupling user-level shared-prefix training with shared-prefix serving for compute-once, decode-many-times ranking. Across industrial and public benchmarks, ReST achieves higher accuracy and scales more consistently along sequence length, depth, and width, where LLM-style Transformer blocks saturate. A one-week online A/B test on a production advertising platform improves online AUC by 1.31% and lifts a core revenue metric by 11.93% within a 50 ms P99 budget; ReST has since been fully deployed in production, showing that behavior-sequence scaling remains a promising, under-exploited axis for production ranking.

Reinforcement learning for user-centric agents is limited by the cost, latency, and risk of collecting online feedback, as well as by the lack of counterfactual comparisons under the same user state. In this paper, we propose World Model-Guided Reinforcement Learning via counterfactual user engagement simulation (WMG-RL), a framework in which a frozen user simulator provides reward supervision before real user exposure. Motivated by language world models, we instantiate the simulator as a User Engagement World Model (UEWM), which treats a recommended item as the agent action and the user's heterogeneous feedback as the environment observation. Rather than learning one fixed environment transition, UEWM learns to infer user-specific dynamics from engagement history and apply them to candidate items. In WMG-RL, a downstream policy proposes multiple candidate items for the same history; UEWM predicts the corresponding engagement feedback in parallel; and the simulated feedback is converted into dense rewards for policy optimization. Experiments show that UEWM provides reliable and transferable reward signals across domains, and that WMG-RL enables a compact 1.7B student policy to match or surpass much larger LLMs on downstream recommendation tasks.

Price extraction from websites is a key task for market monitoring, price comparison, and business analytics in e-commerce. Existing approaches can be broadly divided into four groups, and understanding their trade-offs in accuracy and scalability is essential for selecting suitable extraction strategies. Classical methods rely on manually written wrappers and rule induction from labeled pages, offering high accuracy but adapting poorly to structural changes and requiring considerable maintenance effort. Browser-based methods, using tools such as Selenium and Puppeteer, handle dynamic JavaScript content but consume large computational resources and scale poorly. Browserless approaches retrieve HTML directly via HTTP requests, offering significant gains in speed and cost, but rely on rules calibrated for specific sites. Methods based on machine learning and large language models offer adaptability but require training data and substantial computation. Our main contribution is an adaptive browserless price extraction system that improves robustness to structural differences between websites. We implemented a baseline architecture combining HTML page fragmentation with syntactic, semantic, and frequency rules, and extended it in two ways: a Bayesian approach that dynamically updates rule weights, and a genetic algorithm that optimizes the system's global parameters. This hybrid scheme increased precision from 77.2% to 87.3% and reduced average per-page processing time by approximately 14% relative to the baseline, confirming it as a competitive alternative to manually tuned browserless solutions and to more resource-intensive browser- or LLM-based methods, offering high extraction accuracy at low computational cost.

Industrial recommender systems typically rely on cascaded retrieval, pre-ranking, ranking, and reranking stages, whose separately optimized models limit scaling, fragment decision making, and lack semantic knowledge and reasoning. We present TGR (Tencent Generative Recommendation), an industrial framework that advances recommendation toward the generative paradigm along three coupled directions. TGR-GenRank upgrades ranking through CCFormer, which combines unified feature tokenization, a scalable Transformer backbone, feature-field separated cross attention, subspace token mixing, and hierarchical sequence compression while retaining per-item multi-task outputs. TGR-GenRec explores end-to-end generation under two paradigms: BARGE bridges item-boundary loss and semantic drift in hierarchical semantic-ID generation through item context-aware attention, hierarchical path reranking, and orthogonal dual-path decoding; HiGR performs whole-slate generation with prefix-structured semantic IDs, coarse-to-fine decoding, and listwise multi-objective alignment. TGR-Reason injects offline-generated semantic-ID reason tokens into online decoding, providing reasoning without request-time rollout. TGR is deployed across Tencent production surfaces serving hundreds of millions of users. CCFormer delivers significant gains in five A/B-tested scenarios and is fully launched in two, including +3.57% CTR and +1.71% advertising revenue. BARGE improves Hit@5 by 10.2-16.9% and yields +0.60% CTR and +1.70% reading time after full rollout. HiGR improves offline slate quality by 15.9-21.3% with a 5x inference speedup and achieves up to +1.22% watch time and +1.73% video views. TGR-Reason raises cold-start new-user Hit@1 by 477.8% and delivers +1.75% effective consumption and +13.09% new-user exposure-to-conversion online.

Interactions with cold items negatively impact real-time personalization of ID-based recommender systems. This is because the use of such interactions degrades user preference estimates, whereas excluding cold items from the user profile prevents real-time recommendation updates. In industrial scenarios, one heuristic often applied to address this shortcoming at inference time is to replace, i.e., "swap", cold-start items by their most similar "warm" neighbor, where similarity is inferred from the items' side information. In this paper, we demonstrate that sequential models, most often used for real-time personalization, are not robust to such swaps, and propose SwapRec, an approach to address this issue. SwapRec relies on using the same swap heuristics already at training time. We apply SwapRec to state-of-the-art models for sequential recommendation and analyze its impact by means of quantitative experiments in three recommendation domains (online shopping, movie, music). The experimental results show that, irrespective of the underlying sequential architecture, our easy-to-implement SwapRec approach allows for substantially more accurate recommendations when in presence of interactions with cold items, simultaneously leading to a larger percentage of cold items in the recommendation lists.

Customer-service QA in an AI contact center (AICC) runs under deployment constraints that benchmark QA misses: tight voice-hotline latency and a high cost for unsupported or wrong automatic answers. We deploy a system that answers only from a closed set of verified QA units: it returns a retrieved unit verbatim, or routes to clarify, abstain, or handoff. The index is enriched offline by staged linguistic seeding (SLS): a human authors a per-unit world-grounded slot recipe, gpt-4.1-mini renders it into variants, and a light human gate filters them. One methodology is reused across both domains, so inference stays a single retrieval pass with no query-time generation. On held-out query variants from two industrial domains, SLS lifts hybrid R@1 to 0.881/0.930 (+0.27/+0.34), with gains across all five retrievers tested. At the same gpt-4.1-mini generation budget, SLS beats doc2query by +0.20/+0.32, while cross-provenance evaluation provides additional evidence of transfer across generated-query distributions. Verified-unit answering also removes free-form generation's unsupported-content surface (7-13% versus approximately 0%). We report this as an application study, including negative results.

As far-right actors increasingly exploit online platforms to disseminate ideology and mobilize supporters, civil society organizations (CSOs) play a vital yet underrecognized role in monitoring antidemocratic dynamics online. Unlike fact-checkers or content moderators, CSOs engage in long-term, contextualized analysis, often in resource-constrained settings and under precarious conditions. Despite their critical societal role, CSOs face significant barriers to adopting or co-developing technical solutions, including legal uncertainty, limited platform access, and chronic underfunding. Existing research and tool development efforts have largely overlooked these actors in favor of more institutionally embedded stakeholders. This paper addresses this gap through a qualitative study with 15 practitioners from 12 Germany-based CSOs engaged in online monitoring, positioning them as key yet overlooked stakeholders in the governance of digital spaces. We explore their current practices, challenges, and expectations regarding technological support. Our findings show that monitoring remains largely manual due to the lack of tailored tools, with enhanced search capabilities emerging as the most pressing technical need. While participants express openness to AI-supported features such as media processing and content discovery, many remain skeptical of automated classification, citing concerns around trust, legal usability, and professional credibility. Grounded in these findings, we introduce a conceptual monitoring workflow and describe its implementation in an open-source Telegram monitoring prototype designed to flexibly support diverse monitoring goals. We outline concrete design, policy, and research recommendatios, and introduce the manual labor trap as an empirically grounded concept that explains why monitoring CSOs tend to remain locked into labor-intensive, low-capacity arrangements.

Large vision-language models used as listwise rerankers must jointly process visual tokens from tens of candidates per query, making token pruning essential for practical deployment. Existing pruning methods retain tokens by attention saliency, yet we show that saliency is systematically misaligned with ranking contribution: visually prominent tokens often capture order-neutral patterns shared across candidates. This mismatch is layer-dependent: saliency becomes informative only where attention is concentrated, and normalized attention entropy diagnoses the reliability shift (Pearson r=0.87). We propose RaDiCal (Rank-Discriminative Calibration), a training-free framework that uses normalized attention entropy to decide when saliency can be trusted, fusing it with an attention-free rank-discriminative prior and selecting pruning layers from the same trust landscape. Across three retrieval benchmarks and multiple VLM architectures, RaDiCal matches Dense MRR@10 on Flickr30K and surpasses it on MSCOCO at a 20% token budget, ranks first among all pruning methods on FashionIQ, and holds within 1.2 pp on Flickr30K and MSCOCO at 10% retention. It cuts FLOPs by 39–45% and delivers 1.28–1.45$\times$ measured speedups across two VLM architectures without dataset-specific retuning.

Retrieval is the first stage of modern search and advertising systems, selecting a candidate set from a large item universe for downstream ranking and auction. Recent work increasingly leverages LLMs to improve retrieval through query expansion, data synthesis, and retrieval-feedback training. However, the generative component is typically used for query-side augmentation, while final matching is still delegated to a downstream retriever. We introduce CoGR, a retrieval framework that instead trains LLMs to directly construct retrieval representations on both query and item sides. Each generator produces a compact set of keywords, which are matched directly through an inverted index, preserving compatibility with existing keyword-based retrieval infrastructure. CoGR uses a two-stage training pipeline. Supervised fine-tuning first establishes an aligned keyword space, after which co-evolving reinforcement learning alternately optimizes the query- and item-side generators with GRPO against the opposite side's frozen index. Both sides optimize the same query-to-item retrieval $F_1$ objective: the query side receives retrieval $F_1$ directly, while the item side receives a counterfactual marginal reward measuring the change in query-side $F_1$ caused by its generated keywords. Across 10 representative sparse, dense, and generative baselines, CoGR achieves the best performance on both an internal APP Marketplace dataset and the public WANDS benchmark, improving $F_1$ over the strongest baseline by $10.9\%$ and $36.1\%$, respectively. Further analysis shows stable co-evolution and increasingly aligned query–item keyword spaces over training.

We investigate the role of conversational context modeling in user preference tracking for Conversational Recommendation Systems (CRSs). In this regard, we propose DREAMS, a novel tree-structured context modeling framework that explicitly captures user preference evolution throughout multi-turn interactions. DREAMS introduces two specialized node types to support the two fundamental objectives of CRSs: preference elicitation and preference exploitation. Specifically, elicitation nodes leverage Monte Carlo Tree Search (MCTS) to strategically explore conversational actions and infer latent user preferences, while exploitation nodes employ LLM-based refinement to transform the tracked preference state into structured retrieval queries for recommendation. Extensive experiments on benchmark datasets demonstrate the effectiveness of DREAMS and its design.

The growing complexity of cyber-physical attack surfaces in advanced manufacturing has made cyber threat intelligence analysis increasingly difficult. Although large language models and retrieval-augmented generation have improved CTI workflows, text-based approaches remain vulnerable to hallucinations and provide limited support for structured reasoning over interconnected threats. Graph-based RAG reduces some of these limitations, but existing approaches often lack ontology-consistent multi-hop reasoning and transparent evidence tracing across heterogeneous cybersecurity data. This paper proposes a graph-grounded neuro-symbolic framework that integrates ontology-aware symbolic query generation, knowledge graph retrieval, and neural language generation to support accurate and explainable threat analysis across information technology and operational technology environments. The framework adopts a dual-large language model architecture: the first model translates natural-language questions into executable Cypher queries for symbolic graph retrieval, while the second generates answers strictly from the retrieved graph evidence. Experimental evaluation using publicly available cyber threat intelligence benchmarks shows consistent improvements over the published baseline in reasoning accuracy, while also reducing hallucinations, strengthening multi-hop reasoning, and improving robustness to adversarial perturbations. Runtime and explainability analyses further demonstrate that the framework maintains interactive inference performance and exposes graph-grounded reasoning artifacts that allow analysts to inspect and verify each stage of the analysis. Overall, the results highlight the potential of graph-grounded neuro-symbolic reasoning as a scalable, interpretable, and reliable approach to cyber threat intelligence for next-generation Industry 5.0 environments.

Graph-based Retrieval-Augmented Generation (GraphRAG) supports multi-hop reasoning by organizing corpora into structured graphs. However, graph reachability often captures semantic association rather than evidence support, so a reachable passage may still fail to justify a required cross-passage transition. We propose EvLink, an evidence-linking retriever that preserves passages as retrievable evidence units and builds evidence-supported transitions between them. EvLink constructs two types of reliable links: relation-grounded evidence links justified by explicit source relations, and endpoint-alignment links serving as sourcebounded fallbacks. For retrieval, we introduce a two-stage retrieval strategy. First, bounded breadth-first search over source-grounded evidence links recovers bridge passages missed by similarity-based methods. Then, evidenceneed mining with noisy-OR coverage refinement selects a compact, non-redundant evidence set satisfying distinct question facets. Experiments on three multi-hop and two simple QA benchmarks show EvLink consistently outperforms leading GraphRAG baselines with average gains of 2.4 R@5, 1.9 EM, and 2.4 F1

Multimodal models often build on architectures designed for generative vision-language modeling, typically combining separately pretrained vision encoders with causal language models. Visual document retrievers such as ColPali repurpose these models as encoders, carrying over the parameter and compute overhead of a VLM for a non-generative task. We introduce NeoMME, a family of 260M and 800M-parameter Multimodal and Multilingual bidirectional Encoders that process multilingual text and raw image patches in a single bidirectional Transformer encoder. Both models are pretrained from scratch with a masked discrete-diffusion text objective, conditioned on visible image patches for multimodal examples. Both support a 16,384-token context, enough to encode up to two standard 4K UHD images. To demonstrate its downstream capabilities, we fine-tune NeoMME with jointly trained dense and late-interaction heads. On the ViDoRe v3 benchmark, the resulting NeoMME-Retriever 260M outperforms all evaluated models strictly below 800M parameters with 0.523 nDCG@10, while NeoMME-Retriever 800M reaches 0.556. At a matched 2048x2048 image input size on an NVIDIA L40S, NeoMME-260M encodes pages with about 2x the throughput of ColModernVBERT. Hierarchical token pooling and asymmetric quantization compress late-interaction multimodal document embeddings by 255x while preserving over 95% of baseline nDCG@10. We contribute NeoMME to Hugging Face Transformers and release the pretrained backbone and retrieval-compatible checkpoints under Apache 2.0 at https://hf.co/collections/Hcompany/neomme.

Unified recommendation models aim to jointly model non-sequential multi-field features and sequential user behaviors, but existing interaction-centric designs mainly focus on mixing heterogeneous tokens within each layer. We argue that scalable unified recommendation also requires controlling how intent information is carried, filtered, and preserved across stacked blocks. Inspired by flow-based representation dynamics, we introduce feature transport, a view that treats deep unified recommendation as a discrete context-conditioned representation evolution process. We propose CRAFT, a Contextual Residual Adaptive Feature Transport block, which summarizes non-sequential features into a reliability-aware contextual field and uses it to generate residual displacement and memory-preserving signals for intent and sequence representations. In this way, non-sequential context acts as an active controller of representation evolution rather than a passive object of interaction. In the TAAC2026 advertising recommendation competition, CRAFT achieves a test AUC of 0.838090, surpassing the previous leaderboard-best score of 0.83798. Scaling experiments further show that CRAFT benefits from both depth and width expansion: stacking CRAFT to six blocks improves test AUC to 0.838148, while increasing the hidden dimension reaches 0.838106. These results demonstrate the effectiveness, scalability, and generalization potential of the feature transport paradigm. Source code: https://github.com/AshleyLuo001/CRAFT

Existing text-video retrieval datasets primarily consist of short-form clips containing a single dominant event. While suitable for measuring basic vision-language alignment, they are limited in capturing real-world retrieval scenarios, where long-form videos naturally contain multiple semantically distinct events and a single text query may correspond to several non-contiguous temporal segments. To bridge this gap, we introduce MELON, the first large-scale dataset designed to extend text-video retrieval to long-form videos featuring complex, multi-event structures. MELON explicitly annotates multiple event intervals per video along with their corresponding textual descriptions, enabling both training and evaluation of multi-event understanding in long, untrimmed videos. In addition, we propose a multi-event aware loss that encourages models to differentiate between full-event and partial-event matches, yielding substantial improvements in retrieval accuracy. Together, the MELON dataset and our proposed loss establish a robust foundation for expanding text-to-video retrieval to complex long-form scenarios and provide a more realistic evaluation setting for future research in the field.

Recommender systems in two-sided matching platforms are commonly optimized for immediate engagement signals such as click-through rate, reply rate, or the total number of successful matches. However, in real-world marketplaces, maximizing matches alone may be misaligned with business goals such as user churn rate and platform revenue, especially when users with fewer matches are at substantially higher risk of churn. In this paper, we study a real job matching platform and show that users with very few recent matches are indeed much more likely to leave the platform, while additional matches for already successful users provide limited marginal value for retention. Motivated by this empirical finding, we formulate a retention-aware recommendation problem and implement a simple post-processing method that adjusts the baseline match-focused ranking to prevent user churn. Specifically, the implemented method gives a score boost to churn-risk users with the goal of increasing their likelihood of obtaining matches and improving retention. We evaluate this practical approach in an online experiment on a real job-matching platform. The treatment group showed directionally lower user churn than the control group, although the estimated effect was not statistically significant at conventional levels, while company-side churn showed no evidence of deterioration. To our knowledge, this is among the first online experimental studies to investigate retention-focused recommendation in a real reciprocal job-matching platform.

Retrieval-Augmented Generation (RAG) grounds large language models in external corpora, but implicit trust in retrieved documents creates a critical attack surface: PoisonedRAG shows that a handful of crafted passages can dominate dense retrieval and steer generation toward attacker-chosen answers. We present the Tri-Layer Sieve, a middleware defense that sanitizes retrieved evidence through cross-embedding-space clustering with an independent judge model, structural filtering of trigger-payload artifacts, and LLM consistency verification. The design exploits a key weakness of retrieval-stage poisoning: a single document must satisfy one embedding geometry, one internal Trigger-Payload structure, and one generation objective - rarely all three simultaneously, a fragility that persists even against an adaptive attacker who paraphrases around it. On Natural Questions, HotpotQA, and MS-MARCO with Contriever retrieval (k=50), the Sieve reduces black-box Attack Success Rate from 67.0/87.0/64.0% to 3.0/14.0/4.0%, mitigates white-box HotFlip attacks from ~74% to 27.8% on NQ with Layer 3 enabled, and drives poisoned-document MRR to 0.000, while restoring clean accuracy from 13-33% under attack to 58-76%. Under an architecture-aware adversary who paraphrases triggers to evade the structural filter, enabling the consistency layer halves adaptive ASR (32.0% to 15.0% on NQ) while raising clean accuracy by 18 points, at an added latency of ~16-19 s/query under live retrieval.

Embedding models are trained and evaluated as if retrieval were exact; in production they serve behind approximate indexes – HNSW, IVF, product quantization, or the fixed-dimensional encodings (FDEs) of late-interaction models – whose behavior the encoder's benchmarks never see: one modern encoder recovers just 14% of its exact top-10 through its raw FDE index. Such failures surface only after an index is built, and the standard patches – corpus-fitted transforms such as whitening – must be fitted, stored, and refit as the corpus changes, and can silently rewrite what the encoder returns. This paper shows that index behavior is predictable before anything is built, from label-free statistics of the raw embeddings, through a ladder of instruments matched to what each index family consumes: (1) closed-form moment statistics for the fixed-grid quantizers (PQ, FDE); (2) simulation on a synthetic twin corpus – cluster statistics made generative, on which any index, composed production systems included, can be built and tested – for partition indexes; (3) size-extrapolated, lightly calibrated twins for graph indexes at million-document scale. Predictions land within 0.03 of measured recall on an unseen million-document corpus. The same geometry is trainable: targeting the one statistic no post-hoc transform can move – the score margin – lifts recall for every index family at once, at a small measured task cost. The result: index choice, correction pricing, and production recall forecast from one cheap measurement pass, on new corpora and new indexes alike; serving without per-corpus transform machinery, suited to continuously changing corpora; and a recall-compute frontier pushed by adapting encoders to geometry rather than coupling them to any single index.

Online health information seeking is shifting from keyword search, where users consider a ranked list of links, to conversational systems that compose a single answer and curate its citations. Source evaluation therefore passes from user to platform, yet what these systems surface is poorly characterized. We audited three free consumer products (ChatGPT, Perplexity, Google AI Overview) on twenty English mental health questions under two prompt conditions, with a subset of three also translated into six further languages of varying resource tiers. We recorded 15,942 citations across 1,140 responses and 1,713 unique domains, then classified every citation with a nine-category organizational typology applied by a deterministic classifier validated against human coding. Citations were heavily concentrated: the ten most-cited domains accounted for 43.6% of English citations, and government, commercial health, and academic sources were closely matched at roughly 22% each. Platforms differed little in typical citation volume but sharply in consistency and in the source types they favored. Explicitly requesting sources shifted composition only modestly. Non-English queries surfaced fewer citations and were routed to language-appropriate resources at significantly lower rates. We release the typology, classifier, and annotated corpus as reusable instruments for auditing generative health search.

Scientific discovery depends on finding prior literature that shapes what comes next. Existing retrieval systems optimize for relevance and popularity, often favoring central papers over less familiar works that later prove generative. We introduce MUSES, a million-instance benchmark for prospective intellectual-roots retrieval over a fixed 2.33M-paper corpus, with roughly 140K test instances per familiarity tier. To our knowledge, it is the first prospective benchmark at this scale with a shared retrieval task and author-confirmed paper-level root labels. Alongside it, CiteRoots pairs a scalable rhetorical layer over local citation text (LLM judge $\kappa = 0.896$ versus human gold) with a paper-level author-endorsed layer ($n = 1{,}518$ generative-inspiration pairs from 753 focal papers). MUSES organizes difficulty along two axes: a familiarity axis spanning CiteNext, CiteNew, and CiteNew-Isolated, and a functional axis spanning broad citations, rhetorical roots, and author-endorsed roots. Across 9 method classes, a lean multi-centroid retriever built on SPECTER2 is strongest. Hit@100 falls from 0.534 on CiteNext to 0.424 on CiteNew, 0.205 on rhetorical CiteNew, and 0.171 on author-endorsed CiteNew, a $3.1\times$ decline. In a registered eight-lens full-test audit, roughly half of broad-tier test instances remain unsolved at K=1{,}000. Rhetorical role and author endorsement are distinct: the same judge agrees with endorsement at $\kappa = 0.037$. We release MUSES, both CiteRoots layers, and a distilled open companion judge for future work on prospective retrieval and intellectual roots.

Two-sided digital platforms are inherently dynamic: user preferences shift, item popularity evolves, and reviews both reflect and drive these changes. Yet most sequential recommendation systems treat reviews as passive signals for updating user states, leaving two aspects underexplored. First, review generation is nonrandom, depending on evolving latent states of both users and items. Second, reviews can reshape item states, induce spillover across related items, and influence future user decisions. To address these gaps, we propose a two-sided state-space model (TS-SSM) for event-conditioned sequential recommendation. TS-SSM consists of three components: (1) a modality-missing-not-at-random fusion module that encodes review content and informative observation patterns; (2) user-state evolution with temporal variation and local graph message passing that uses related item states to refine user preferences; and (3) item-state evolution with asymmetric carryover of positive and negative review feedback. In experiments across six Amazon categories, TS-SSM increases Recall@20 over BSARec by 14.8%–18.8% and exceeds HM4SR by 11.7% on average. On Goodreads Fantasy, Recall@20 improves HM4SR from .5191 to .5847. Ablations highlight distinct contributions of observation patterns, local propagation, and item dynamics.

BioMedRAG introduced retrieval-augmented generation with a learned chunk scorer for biomedical information extraction. However, it relies on fixed-size chunking which can fragment semantic evidence. We propose a configurable semantic chunking framework that addresses this limitation by combining entity-preserving windows, trigger-centered chunking, proposition-first extraction, tiered trigger prioritization, and hierarchical relation resolution. The framework integrates with BioMedRAG by replacing only the chunk construction stage while preserving the embedding model, learned chunk scorer, generator, and evaluation protocol. We evaluate the framework on biomedical relation extraction benchmarks (GM-CIHT, DDI, ChemProt) and adverse event classification (ADE). On GM-CIHT, the full hybrid configuration achieves 82.6% F1, improving over the fixed-size baseline (74.2% F1) by 8.4 points under our experimental setup. Cross-dataset analysis shows that semantic chunking improves extraction datasets with explicit relation cues, such as GM-CIHT and DDI, while fixed chunking remains competitive or stronger for dense biochemical extraction and binary classification settings such as ChemProt and ADE. By externalizing chunking logic into configuration files, the framework provides an interpretable and adaptable alternative to rigid fixed-size chunking for biomedical RAG pipelines.