arXiv subjects

Mounia Lalmas

Publications and source records attributed to Mounia Lalmas.

At least 19 recordsLinked to original sources

arXiv2026

Robust Fusion of Semantic and Behavioural Signals for LLM Reranking in Personalised Search

Personalised search must satisfy query intent while incorporating user context and historical interactions. LLM-based cross-encoders provide a single reranking interface, but injecting predictive behavioural statistics into their prompts can encourage shortcut learning: reliance on historical signals at the expense of semantic and user-context patterns that generalise to sparse or unseen searches. We study this problem in the personalised search system of a large-scale audio streaming platform using Query Slice Stats (QSS), an interaction-derived behavioural feature summarising historical success for query-candidate pairs. Naive QSS injection improves ranking when the feature is available but reduces robustness when it is removed. We address this with deterministic dual-sample feature-dropout training, which presents each example once with QSS included and once with QSS removed. Offline, QSS injection improves ranking quality by 13.3% when available. Dual-sample training preserves these gains while improving performance under QSS-removed evaluation by 4.0% relative to naive QSS training. In a live online test, both QSS-aware variants improve search success by roughly 2%. The aggregate test does not distinguish dual-sample from features-only training; the cold-start comparison is directionally consistent with the offline results. Paired feature-present and feature-removed training can therefore reduce the tension between exploiting strong behavioural statistics and remaining robust when they are unavailable.

cs.IR↗

arXiv2026

The Disconnect Between Better Descriptive Reasoning Trace Quality and Recommendation Effectiveness

Recent work has focused on improving explicit natural-language descriptive reasoning traces for generative recommendation. This includes systems that augment semantic ID (SID) prediction with chain-of-thought reasoning. However, because SIDs are opaque learned identifiers rather than natural language, they require costly alignment before an LLM can reason over them. This provides a controlled experimental setting in which both item representation (Title vs. SID) and semantic grounding (minimal vs. extensive SID alignment) can be varied independently. We therefore present the first controlled comparison of descriptive reasoning trace quality across semantic IDs and natural-language titles in a 2 x 2 factorial study on three Amazon product domains using a shared Qwen3-1.7B backbone. We find that introducing explicit descriptive reasoning traces reduces traditional offline recommendation effectiveness under standard SFT and RL training, even though natural language titles produce substantially more grounded and interpretable traces. Extensive SID alignment improves descriptive trace quality but not traditional offline recommendation effectiveness, while a richer reward signal partially recovers performance. Overall, our results show that improving descriptive reasoning trace quality is not, by itself, sufficient to consistently improve traditional offline recommendation effectiveness under the training objectives and evaluation protocols studied here.

cs.IR↗

arXiv2026

Do Sequential Recommendation Benchmarks Really Require Higher-Order Sequence Modelling?

Sequential recommenders increasingly use language-model architectures designed to capture complex, context-dependent interactions. Yet it remains unclear whether widely used benchmarks actually require this modelling capacity. We investigate this question using two simple, recency-weighted pairwise probes that do not learn higher-order sequence representations: Sequential Rules (SeqRules) and our Probabilistic Collaborative Transition Model (PCTM). Using the evaluation protocol of eSASRec, at least one probe exceeds our eSASRec reproduction by 15-38% on three Amazon datasets and by 4.4% on MovieLens-1M, but trails it by 27.3% on MovieLens-20M. On the four remaining datasets, at least one probe also outperforms our sampled-softmax SASRec reproduction by 9-28%, suggesting that these widely used benchmarks are poorly suited to measuring gains from higher-order sequence modelling. More broadly, comparing Transformer-based models against strong recency-weighted pairwise probes provides a concrete test of whether a benchmark can meaningfully measure gains from higher-order sequence modelling.

cs.IR↗

arXiv2026

From IR to RecSys: Evaluating LLM-based Judges in Cranfield-style Recommendation Collections

The Cranfield paradigm has long provided reliable, reproducible evaluation in ad hoc retrieval, and recent work has begun extending this framework to recommender systems. A recent development in IR is the use of Large Language Models (LLMs) as automatic relevance judges, showing promising agreement with human assessors. Whether this LLM-judge paradigm---studied predominantly on query--document pairs---transfers to the subjective, profile-driven nature of recommendation remains an open question. This paper bridges the IR and RecSys evaluation traditions by systematically investigating LLM-based judges within a Cranfield-style recommendation collection. Using the ML-32M-ext movie recommendation collection, we first demonstrate that traditional train--test splits yield substantially incomplete relevance labels and unreliable system rankings compared to Cranfield-style pooling. We then assess LLM-judge alignment with human labels, finding that richer item metadata and longer user histories improve agreement, although item-level agreement remains moderate overall. Rankings derived from LLM-judge labels achieve high agreement with human-based rankings (Kendall's tau up to 0.92 for nDCG@100 across 52 system configurations), comparable to values reported for TREC ad hoc retrieval collections. Crucially, LLM-judge recovers system rankings that are distorted under traditional evaluation---correctly identifying systems that are undervalued or overvalued by incomplete labels. An industrial case study in podcast recommendation further demonstrates the practical value of LLM-judge for model selection. Rather than positioning LLM-judges as a replacement for human or interaction-based evaluation, our results support their use as a promising complementary signal: item-level agreement with humans is moderate, yet system-level rankings---which aggregate judgments over many user--item pairs---remain stable.

cs.IR↗

arXiv2026

Hypothesis-Driven Shelf Generation for Personalised Recommendation

Modern recommendation interfaces organise content into shelves: themed rows such as "More of What You Like" or "New Releases for You." In production systems, these shelves are typically defined through hand-crafted templates coupled with dedicated retrieval logic. While effective for broad recommendation intents, this approach does not scale to the long tail of individual taste. We present a content-hypothesis-driven shelf generation system for Spotify Home that replaces fixed templates with natural-language hypotheses describing what a personalised shelf should contain. The system has four stages hypothesis generation, catalogue fulfilment, shelf alignment, and offline serving. This decomposition decouples shelf planning from catalogue fulfilment, supports independent optimisation of planning and retrieval, and enables both constrained generative retrieval over catalogue entities and distillation of frontier LLM behaviour into compact models. Our production pipeline combines hypothesis generation, generative retrieval, candidate selection and shelf alignment, offline LLM-as-a-judge evaluation, and precomputed serving. We describe the end-to-end architecture and evaluate it through offline analyses and an early online evaluation under uniform random exposure on Spotify Home. Results show that hypothesis-driven shelves substantially expand personalised recommendation supply with engagement that varies by content type and is competitive with strong existing shelves in some settings.

cs.IR↗

arXiv2026

Who Are We Recommending To? Recommender Systems in the Agentic Web

For two decades, recommender systems have been designed under the assumption that a human directly consumes each recommendation: receiving, interpreting, and acting upon it. The emergence of AI agents powered by large language models challenges this assumption. In the emerging Agentic Web [ 28 ], autonomous agents increasingly act on behalf of users, e.g., browsing, comparing, negotiating, and executing transactions, raising a central question: who is the receiver of a recommendation? In this position paper, we argue that the recommendation paradigm is undergoing a bifurcation. In delegable contexts, such as routine purchases, travel, and constrained transactional tasks, the primary operational consumer of recommendations is shifting from the human to the agent, requiring new optimization objectives, interaction protocols, and evaluation criteria. In experiential contexts, such as entertainment, art, and other subjective or high-stakes choices, humans remain the final judge of relevance, though agents may assist through pre-filtering and curation. We introduce a delegation spectrum that characterizes recommendation contexts along factors such as preference specifiability, outcome verifiability, and decision stakes, and we outline a research agenda spanning agent preference modeling, dual-audience optimization, and the emerging agent attention economy. We further discuss the implications of this shift for the design and evaluation of recommender systems

cs.IR↗

arXiv2026

As It Was: Aligning LLM Search Evaluation with Historical User Preferences

Large-scale search systems evolve faster than human quality assurance can scale, especially for long-tail intents and multilingual queries. LLM-as-a-judge approaches provide a scalable alternative for evaluating the relevance of search engine result pages (SERPs), but judgments based solely on semantic similarity or world knowledge can drift from actual user preferences, particularly for ambiguous queries. We introduce a behavior-grounded LLM judge that augments each SERP item with a lightweight and auditable behavioral prior in the form of a Query-Relevance-Impressions (QRI) card. Each card summarizes how users have historically interacted with similar queries and results, providing compact empirical evidence that the judge can cite to resolve ambiguity and make more consistent relevance judgments while still relying on semantic reasoning. In a large-scale music search evaluation at Spotify, using relevance estimates derived from historical user interactions across 6,000 recomposed SERPs, the behavior-grounded judge achieves stronger alignment with user preferences, improving Spearman rank correlation by approximately 5% overall and yielding a 91% relative improvement on disagreement cases. On a multilingual human-judged dataset spanning five languages, grounding further increases correlation with human relevance judgments by 15%. Importantly, when evaluated against outcomes from a live A/B test, the grounded judge shows consistently higher alignment with the observed winning model. While absolute alignment remains moderate, these findings demonstrate that lightweight behavioral grounding can improve the reliability and practical usefulness of LLM-based evaluation in real-world search systems.

cs.IR↗

arXiv2026

Stochastic Primal-Dual Decoding for Multiobjective Generative Recommender Systems

Recent advances in recommender systems (RS) have shown substantial performance gains through generative modelling. In practice, recommendation often involves constructing slates -- ordered lists of items -- that must satisfy multiple objectives beyond relevance, such as constraints defined over item attributes or fairness constraints. Existing multiobjective approaches either rely on post-processing techniques designed for non-generative settings, or incorporate auxiliary objectives directly into model training. The former does not explicitly account for the sequential nature of generative RS, while the latter is often impractical in large-scale systems. We propose a lightweight, inference-time decoding layer that augments autoregressive generative RS to support multiobjective slate generation without modifying or retraining the underlying model. We formulate decoding as an online constrained optimisation problem, where items are selected sequentially, and trade-offs between relevance and auxiliary objectives are adjusted dynamically based on the remaining constraint slack, i.e., how much of each objective remains to be satisfied. This is implemented via a stochastic primal-dual approximation scheme that balances relevance and auxiliary objectives during generation. We provide theoretical guarantees on constraint violation and regret, and evaluate the proposed approach through extensive offline experiments and a large-scale online A/B experiment in a real-world recommender system. Our results show consistent improvements in multiobjective trade-offs, including a +1.8\% gain in the auxiliary objectives achieved at zero cost to user satisfaction.

cs.AI↗

arXiv2026

Primal-Dual Guided Decoding for Constrained Discrete Diffusion

Discrete diffusion models generate structured sequences by progressively unmasking tokens, but enforcing global property constraints during generation remains an open challenge. We propose primal-dual guided decoding, an inference-time method that formulates constrained generation as a KL-regularised optimisation problem and solves it online via adaptive Lagrangian multipliers. At each denoising step, the method modifies token logits through an additive, constraint-dependent bias, with multipliers updated by mirror descent based on constraint violation. The bias arises as the optimal KL-regularised projection of the constraint, so the constrained distribution remains as close as possible to the model's unconstrained distribution while still satisfying the constraint. The method requires no retraining and no additional model evaluations beyond standard sampling, supports multiple simultaneous constraints, and provides formal bounds on constraint violation. We evaluate our approach on topical text generation, molecular design, and music playlist generation, showing that a single algorithm instantiated via domain-specific scoring functions improves constraint satisfaction while preserving relevant domain-specific quality metrics.

cs.AI↗

arXiv2026

Efficient Dataset Selection for Continual Adaptation of Generative Recommenders

Recommendation systems must continuously adapt to evolving user behavior, yet the volume of data generated in large-scale streaming environments makes frequent full retraining impractical. This work investigates how targeted data selection can mitigate performance degradation caused by temporal distributional drift while maintaining scalability. We evaluate a range of representation choices and sampling strategies for curating small but informative subsets of user interaction data. Our results demonstrate that gradient-based representations, coupled with distribution-matching, improve downstream model performance, achieving training efficiency gains while preserving robustness to drift. These findings highlight data curation as a practical mechanism for scalable monitoring and adaptive model updates in production-scale recommendation systems.

cs.IR↗

arXiv2026

A Unified Language Model for Large Scale Search, Recommendation, and Reasoning

LLMs are increasingly applied to recommendation, retrieval, and reasoning, yet deploying a single end-to-end model that can jointly support these behaviors over large, heterogeneous catalogs remains challenging. Such systems must generate unambiguous references to real items, handle multiple entity types, and operate under strict latency and reliability constraints requirements that are difficult to satisfy with text-only generation. While tool-augmented recommender systems address parts of this problem, they introduce orchestration complexity and limit end-to-end optimization. We view this setting as an instance of a broader research problem: how to adapt LLMs to reason jointly over multiple-domain entities, users, and language in a fully self-contained manner. To this end, we introduce NEO, a framework that adapts a pre-trained decoder-only LLM into a tool-free, catalog-grounded generator. NEO represents items as SIDs and trains a single model to interleave natural language and typed item identifiers within a shared sequence. Text prompts control the task, target entity type, and output format (IDs, text, or mixed), while constrained decoding guarantees catalog-valid item generation without restricting free-form text. We refer to this instruction-conditioned controllability as language-steerability. We treat SIDs as a distinct modality and study design choices for integrating discrete entity representations into LLMs via staged alignment and instruction tuning. We evaluate NEO at scale on a real-world catalog of over 10M items across multiple media types and discovery tasks, including recommendation, search, and user understanding. In offline experiments, NEO consistently outperforms strong task-specific baselines and exhibits cross-task transfer, demonstrating a practical path toward consolidating large-scale discovery capabilities into a single language-steerable generative model.

cs.IR↗

arXiv2026

Deploying Semantic ID-based Generative Retrieval for Large-Scale Podcast Discovery at Spotify

Podcast listening is often grounded in a set of favorite shows, while listener intent can evolve over time. This combination of stable preferences and changing intent motivates recommendation approaches that support both familiarity and exploration. Traditional recommender systems typically emphasize long-term interaction patterns, and are less explicitly designed to incorporate rich contextual signals or flexible, intent-aware discovery objectives. In this setting, models that can jointly reason over semantics, context, and user state offer a promising direction. Large Language Models (LLMs) provide strong semantic reasoning and contextual conditioning for discovery-oriented recommendation, but deploying them in production introduces challenges in catalog grounding, user-level personalization, and latency-critical serving. We address these challenges with GLIDE, a production-scale generative recommender for podcast discovery at Spotify. GLIDE formulates recommendation as an instruction-following task over a discretized catalog using Semantic IDs, enabling grounded generation over a large inventory. The model conditions on recent listening history and lightweight user context, while injecting long-term user embeddings as soft prompts to capture stable preferences under strict inference constraints. We evaluate GLIDE using offline retrieval metrics, human judgments, and LLM-based evaluation, and validate its impact through large-scale online A/B testing. Across experiments involving millions of users, GLIDE increases non-habitual podcast streaming on Spotify home surface by up to 5.4% and new-show discovery by up to 14.3%, while meeting production cost and latency constraints.

cs.IR↗

arXiv2025

Calibrated Recommendations with Contextual Bandits

Spotify's Home page features a variety of content types, including music, podcasts, and audiobooks. However, historical data is heavily skewed toward music, making it challenging to deliver a balanced and personalized content mix. Moreover, users' preference towards different content types may vary depending on the time of day, the day of week, or even the device they use. We propose a calibration method that leverages contextual bandits to dynamically learn each user's optimal content type distribution based on their context and preferences. Unlike traditional calibration methods that rely on historical averages, our approach boosts engagement by adapting to how users interests in different content types varies across contexts. Both offline and online results demonstrate improved precision and user engagement with the Spotify Home page, in particular with under-represented content types such as podcasts.

cs.LG↗

arXiv2025

Prompt-to-Slate: Diffusion Models for Prompt-Conditioned Slate Generation

Slate generation is a common task in streaming and e-commerce platforms, where multiple items are presented together as a list or ``slate''. Traditional systems focus mostly on item-level ranking and often fail to capture the coherence of the slate as a whole. A key challenge lies in the combinatorial nature of selecting multiple items jointly. To manage this, conventional approaches often assume users interact with only one item at a time, assumption that breaks down when items are meant to be consumed together. In this paper, we introduce DMSG, a generative framework based on diffusion models for prompt-conditioned slate generation. DMSG learns high-dimensional structural patterns and generates coherent, diverse slates directly from natural language prompts. Unlike retrieval-based or autoregressive models, DMSG models the joint distribution over slates, enabling greater flexibility and diversity. We evaluate DMSG in two key domains: music playlist generation and e-commerce bundle creation. In both cases, DMSG produces high-quality slates from textual prompts without explicit personalization signals. Offline and online results show that DMSG outperforms strong baselines in both relevance and diversity, offering a scalable, low-latency solution for prompt-driven recommendation. A live A/B test on a production playlist system further demonstrates increased user engagement and content diversity.

cs.IR↗

arXiv2025

Describe What You See with Multimodal Large Language Models to Enhance Video Recommendations

Existing video recommender systems rely primarily on user-defined metadata or on low-level visual and acoustic signals extracted by specialised encoders. These low-level features describe what appears on the screen but miss deeper semantics such as intent, humour, and world knowledge that make clips resonate with viewers. For example, is a 30-second clip simply a singer on a rooftop, or an ironic parody filmed amid the fairy chimneys of Cappadocia, Turkey? Such distinctions are critical to personalised recommendations yet remain invisible to traditional encoding pipelines. In this paper, we introduce a simple, recommendation system-agnostic zero-finetuning framework that injects high-level semantics into the recommendation pipeline by prompting an off-the-shelf Multimodal Large Language Model (MLLM) to summarise each clip into a rich natural-language description (e.g. "a superhero parody with slapstick fights and orchestral stabs"), bridging the gap between raw content and user intent. We use MLLM output with a state-of-the-art text encoder and feed it into standard collaborative, content-based, and generative recommenders. On the MicroLens-100K dataset, which emulates user interactions with TikTok-style videos, our framework consistently surpasses conventional video, audio, and metadata features in five representative models. Our findings highlight the promise of leveraging MLLMs as on-the-fly knowledge extractors to build more intent-aware video recommenders.

cs.IR↗

arXiv2025

Evaluating Podcast Recommendations with Profile-Aware LLM-as-a-Judge

Evaluating personalized recommendations remains a central challenge, especially in long-form audio domains like podcasts, where traditional offline metrics suffer from exposure bias and online methods such as A/B testing are costly and operationally constrained. In this paper, we propose a novel framework that leverages Large Language Models (LLMs) as offline judges to assess the quality of podcast recommendations in a scalable and interpretable manner. Our two-stage profile-aware approach first constructs natural-language user profiles distilled from 90 days of listening history. These profiles summarize both topical interests and behavioral patterns, serving as compact, interpretable representations of user preferences. Rather than prompting the LLM with raw data, we use these profiles to provide high-level, semantically rich context-enabling the LLM to reason more effectively about alignment between a user's interests and recommended episodes. This reduces input complexity and improves interpretability. The LLM is then prompted to deliver fine-grained pointwise and pairwise judgments based on the profile-episode match. In a controlled study with 47 participants, our profile-aware judge matched human judgments with high fidelity and outperformed or matched a variant using raw listening histories. The framework enables efficient, profile-aware evaluation for iterative testing and model selection in recommender systems.

cs.IR↗

arXiv2025

NeurIPS 2025 E2LM Competition : Early Training Evaluation of Language Models

Existing benchmarks have proven effective for assessing the performance of fully trained large language models. However, we find striking differences in the early training stages of small models, where benchmarks often fail to provide meaningful or discriminative signals. To explore how these differences arise, this competition tackles the challenge of designing scientific knowledge evaluation tasks specifically tailored for measuring early training progress of language models. Participants are invited to develop novel evaluation methodologies or adapt existing benchmarks to better capture performance differences among language models. To support this effort, we provide three pre-trained small models (0.5B, 1B, and 3B parameters), along with intermediate checkpoints sampled during training up to 200B tokens. All experiments and development work can be run on widely available free cloud-based GPU platforms, making participation accessible to researchers with limited computational resources. Submissions will be evaluated based on three criteria: the quality of the performance signal they produce, the consistency of model rankings at 1 trillion tokens of training, and their relevance to the scientific knowledge domain. By promoting the design of tailored evaluation strategies for early training, this competition aims to attract a broad range of participants from various disciplines, including those who may not be machine learning experts or have access to dedicated GPU resources. Ultimately, this initiative seeks to make foundational LLM research more systematic and benchmark-informed from the earliest phases of model development.

cs.AI↗

arXiv2025

Text2Tracks: Prompt-based Music Recommendation via Generative Retrieval

In recent years, Large Language Models (LLMs) have enabled users to provide highly specific music recommendation requests using natural language prompts (e.g. "Can you recommend some old classics for slow dancing?"). In this setup, the recommended tracks are predicted by the LLM in an autoregressive way, i.e. the LLM generates the track titles one token at a time. While intuitive, this approach has several limitation. First, it is based on a general purpose tokenization that is optimized for words rather than for track titles. Second, it necessitates an additional entity resolution layer that matches the track title to the actual track identifier. Third, the number of decoding steps scales linearly with the length of the track title, slowing down inference. In this paper, we propose to address the task of prompt-based music recommendation as a generative retrieval task. Within this setting, we introduce novel, effective, and efficient representations of track identifiers that significantly outperform commonly used strategies. We introduce Text2Tracks, a generative retrieval model that learns a mapping from a user's music recommendation prompt to the relevant track IDs directly. Through an offline evaluation on a dataset of playlists with language inputs, we find that (1) the strategy to create IDs for music tracks is the most important factor for the effectiveness of Text2Tracks and semantic IDs significantly outperform commonly used strategies that rely on song titles as identifiers (2) provided with the right choice of track identifiers, Text2Tracks outperforms sparse and dense retrieval solutions trained to retrieve tracks from language prompts.

cs.IR↗