Jul 2026· Annual International ACM SIGIR Conference on Research and Development in Information Retrieval· 0 citations· 33 references
Computer Science
TL;DR
A controlled SASRec variant is studied that replaces learned attention weights with uniform aggregation and is trained under an otherwise identical block structure and training recipe, and this modification often yields performance comparable to the original model, with clear dataset-dependent exceptions.
Abstract
Causal self-attention models such as SASRec are widely used in sequential recommendation, where learned attention weights are often assumed to provide crucial importance weighting over past interactions. Yet it is unclear when predictive performance truly depends on such non-uniform weighting. We study a controlled SASRec variant that replaces learned attention weights with uniform aggregation and is trained under an otherwise identical block structure and training recipe. Across fourteen benchmark datasets, this modification often yields performance comparable to the original model, with clear dataset-dependent exceptions. To explain this heterogeneity, we introduce a stage-wise norm-based decomposition that quantifies self-preserving vs. cross-position mixing within attention blocks. Across datasets, we find distinct regimes: low mixing yields robustness to uniformization; higher mixing tends to coincide with sensitivity, while some datasets exhibit substantial mixing without dependence on learned weighting. Our results provide a practical diagnostic for identifying when attention weighting is functionally utilized in sequential recommendation. The code is available at: https://github.com/keito0329/revisiting-sasrec.
Large language models produce semantically rich embeddings, yet direct employment of generic LLM embeddings fails to satisfy news recommendation demands due to inherent semantic mismatches with task targets. Existing fine-tuning methods including LoRA can narrow such gaps but bring prohibitive computational overhead, restricting real-world deployment. This work proposes lightweight task-adaptive modules (TAMs). It keeps LLM parameters fixed and transforms offline embeddings into task-specialized representations without full-model backpropagation, drastically cutting training costs. Evaluated on MIND benchmarks across ten mainstream recommendation architectures, TAMs achieves comparable accuracy to LoRA, with computational cost reduced to 1/10 of LoRA’s level. TAMs outperform GloVe-based models by 1.3–12.4% in AUC and scale effectively to MINDlarge. Ablation experiments confirm that the nonlinear projection is pivotal to performance improvement, and statistical validation across three random seeds confirms result robustness. This paradigm provides an efficient low-cost solution for LLM-based news recommendation under resource constraints.
Han Wei, Tong Niu, Sisi Peng et al.· Applied Sciences· 1 citation
Cross-domain recommendation is a well-known technique for improving recommendations in a target domain, especially under sparse data or cold-start conditions. A common strategy is to train user embeddings separately in the source and target domains and learn a transfer function between them. In contrast, we propose SetCDR, which constructs more effective user representations in the target domain by directly incorporating each user's source and target history. We additionally introduce a lightweight domain indicator that preserves data–domain relational information. These histories, composed of item–rating pairs, are represented as variable-length sets and processed using a permutation-invariant neural architecture. This differs from conventional neural networks, which do not naturally handle unordered inputs, and is well suited to recommender systems where user history sizes vary greatly among users. The use of a permutation-invariant architecture ensures consistent embeddings regardless of input order, improving robustness to real-world variability and training efficiency. We demonstrate SetCDR in two forms: a simple sum-pooling method and an extended multihead attention–based method that captures more complex dependencies within user histories. Moreover, because SetCDR operates directly on sets of user history records, it provides a natural way to examine how histories influence user representations. Finally, SetCDR adapts immediately to new interactions without additional retraining, enabling on-the-fly performance improvement. Experimental results across multiple cross-domain benchmarks confirm that SetCDR consistently outperforms strong baselines in recommendation quality.
K. Kim, James Geraci· Annual International ACM SIG...· 0 citations
Understanding which parameters are influential in Large Language Models (LLMs) is central to improving their efficiency, reliability, and interpretability. We introduce Weight-Adjusted Gradients (WAG), a simple yet effective approach for estimating parameter importance that explicitly captures the interaction between model weights and first-order gradient information and identifies parameters that disproportionately influence model behavior, such as those responsible for collapse phenomena in LLMs. Across a range of models and settings, we show that WAG surfaces a tiny but critical subset of parameters whose modification leads to dramatic degradation in performance, a failure mode that existing importance metrics overlook. These findings reveal a previously underexplored interplay between weights and gradients, suggesting that parameter importance cannot be fully understood through either signal alone. The surprising effectiveness of WAG points to fundamental structural properties of trained networks and motivates new open questions about the role of zeroth-order and first-order information in deep learning. We demonstrate the practical utility of WAG across multiple applications, including expert allocation in mixture-of-expert architectures, parameter-specific unlearning, mixed-precision quantization, and layer selection for knowledge editing. Our results position WAG as a unified approach for analyzing, debugging, and controlling LLMs, and opens new directions for principled model-level interpretation.
Sequential recommenders are commonly trained with one-hot next-item labels under a causal (prefix-only) objective aligned with inference. While deployment-compatible, this supervision offers little insight into relative preferences among non-target items. Yet logged interaction sequences contain an additional supervisory source: interactions following the target often reveal how user intent evolves, making the target easier to interpret. We treat these future interactions as training-only privileged information, available during learning but not at inference. This raises a natural question: can future interactions provide richer supervision while keeping training aligned with inference-time prediction? We propose Privileged Self-Distillation (PSD), a framework that separates learning-time information from inference-time information. PSD applies two attention masks to the same backbone: a future-aware view yields a privileged teacher distribution conditioned on past and future interactions, while a prefix-only view yields the student distribution used for deployment. Distilling the privileged distribution converts future interactions into training-only supervision rather than inference-time inputs. Since both views share a backbone, the teacher's advantage is purely informational, not architectural, removing the need for a separately pretrained teacher and letting its supervision adapt as the student evolves. PSD further uses an advantage-reachability gate to focus distillation on teacher signals likely supported by the observed prefix, along with a momentum-averaged teacher for stable targets. The framework is optimized end-to-end in a single stage, leaving the deployed model and inference cost unchanged. Experiments across public benchmarks and diverse backbones show consistent improvements.
Jiakai Tang, Yang Zhang, See-Kiong Ng et al.· 0 citations
Multi-armed bandit algorithms, especially Thompson sampling, are widely used in online recommendation. Despite their ability to adapt from online feedback, these methods often suffer from cold-start limitations when newly introduced arms have little or no interaction history. In our setting, the candidate arms are user-generated textual comments, whose semantic content can reveal a title's appeal before sufficient interaction feedback is available. We therefore use large language models (LLMs) to extract semantic signals from comment text and convert them into informative Bayesian priors that warm-start Thompson sampling under sparse early-stage feedback. To account for aggregate segment-level differences in response patterns, we maintain and update posteriors separately for each gender-age segment. In a real-world online A/B/C test, we compare a uniform prior with two LLM-based designs: a Gender Prior for demographic-affinity cues and a Content Prior for title-specific identity cues. The results show that LLM-based priors are most beneficial in sparse-feedback regimes -- with the largest gains emerging once a small amount of interaction evidence has accumulated -- and that prior design leads to distinct funnel-level effects. We further analyze prior-reward alignment and demographic heterogeneity, finding that click-oriented alignment is strongest for the Gender Prior and that treatment effects vary substantially across demographic segments. These findings suggest that LLM-derived priors can serve as a practical warm-start mechanism for text-rich bandit recommendation, while also revealing deployment trade-offs.
E. Lee, Oseong Choi, Byungsoo Kang et al.· 0 citations
High task performance does not show whether a model retains prediction-relevant structural information in its internal representation. Temporal graph models, for example, can achieve high future-link AUC while basic graph statistics remain difficult to recover from the same representation. We identify one source of this gap in the weighted averaging used by standard attention: when an evidence pattern is repeated, the numerator and denominator grow at the same rate, so inputs with different amounts of accumulated evidence can produce the same aggregate. We propose Mass-Aware Attention (MAA), which generalizes standard L1 normalization to an Lp family. Under repetition, MAA makes the numerator and denominator scale at different rates, retaining the effective number of contributing inputs in the representation magnitude. It adds no supervision, parameters, hidden dimensions, or explicit count features, and recovers standard attention at p=1. Across four continuous-time dynamic graph models and three datasets, MAA improves future-link AUC in 11 of 12 model-dataset cells. Linear recovery from the same hidden representation increases by 4.49% on average, and preferential-attachment recovery improves in all 12 cells after family-wise correction. We also observe consistent evidence in marked temporal point processes, temporal knowledge graphs, retrieval-augmented generation, and spatio-temporal point processes. Information accessibility and task utility remain distinct: NLL improves in MTPP, ranking is largely preserved in TKG, additional information in RAG does not improve the diagnostic head, and downstream LayerNorm can erase the signal in STPP. These results position MAA as a general normalization principle for improving predictor-facing representation informativeness by controlling repetition invariance in standard attention.