MOSAIC: Multi-Domain Orthogonal Session Adaptive Intent Capture for Prescient Recommendations

Abderaouf Bahi; Mourad Boughaba; Ibtissem Gasmi; Warda Deghmane; Amel Ourici

cs.IRApr 11, 2026

MOSAIC: Multi-Domain Orthogonal Session Adaptive Intent Capture for Prescient Recommendations

Authors: Abderaouf Bahi, Mourad Boughaba, Ibtissem Gasmi, Warda Deghmane, Amel Ourici

Organizations: Computer Science and Applied Mathematics Laboratory (LIMA), Faculty of Science and Technology, Chadli Bendjedid University, P.O. Box 73, El Tarf 36000, Algeria · Electro-Mechanical Systems Laboratory (LSEM), Faculty of Technology, Badji Mokhtar University, P.O. Box 12, Annaba 23000, Algeria · Mathematical Modeling and Numerical Simulation Laboratory (LAM2SIN), Faculty of Technology, Badji Mokhtar University, P.O. Box 12, Annaba 23000, Algeria

Abstract

Capturing user intent across heterogeneous behavioral domains stands as a fundamental challenge in session-based recommender systems. Yet, existing multi-domain approaches frequently fail to isolate the distinct contribution of cross-domain interactions from those arising within individual domains, limiting their ability to build rich and transferable user representations. In this work, we propose MOSAIC, a Multi-Domain Orthogonal Session Adaptive Intent Capture framework that explicitly factorizes user preferences into three orthogonal components: domain-specific, domain-common, and cross-sequence-exclusive representations. Our approach employs a triple-encoder architecture, where each encoder is dedicated to one preference type, enforced through domain masking objectives and adversarial training via a gradient reversal layer. Representational alignment and mutual independence constraints are jointly optimized to ensure clean preference separation. Additionally, a dynamic gating mechanism modulates the relative contribution of each component at every timestep, yielding a unified and temporally adaptive session-level user representation. We conduct extensive experiments on two large-scale real-world benchmarks spanning multiple domains and interaction types. The ablation study validates that each component domain-specific encoding, domain-common modeling, cross-sequence representation, and dynamic gating contributes meaningfully to the overall performance. Experimental results demonstrate that MOSAIC consistently outperforms state-of-the-art baselines in recommendation accuracy, while simultaneously providing interpretable insights into the interplay between domain-specific and cross-domain preference signals. These findings highlight the potential of orthogonal preference decomposition as a principled strategy for next-generation multi-domain recommender systems.

Figures & tables

Panel of Table 1 — Table 1: Summary of related work

Dataset	Domain	#Users	#Items	#Interactions	Avg. len.	Density
Amazon Movie–Book	Movie	18,032	64,591	1,041,294	57.7	0.089%
Amazon Movie–Book	Book	18,032	253,841	1,584,033	87.8	0.034%
Amazon Movie–Music	Movie	10,547	50,762	681,430	64.6	0.127%
Amazon Movie–Music	Music	10,547	73,839	456,872	43.3	0.059%
Douban Movie–Book–Music	Movie	2,712	33,519	1,278,401	471.3	1.407%
	Book	2,712	22,347	792,062	292.0	1.308%

Table 2: Dataset statistics after preprocessing

Panel of Figure 4 — Fig. 2 : Baselines classification by category

	Amazon Movie–Book (Movie)						Amazon Movie–Music (Movie)						Douban Movie–Book (Movie)
Method	HR@5	HR@10	HR@20	N@5	N@10	N@20	HR@5	HR@10	HR@20	N@5	N@10	N@20	HR@5	HR@10	HR@20	N@5	N@10	N@20
GRU4Rec	0.0412	0.0683	0.1042	0.0271	0.0361	0.0470	0.0387	0.0641	0.0973	0.0254	0.0339	0.0440	0.0631	0.1023	0.1561	0.0418	0.0553	0.0719
SASRec	0.0524	0.0851	0.1293	0.0347	0.0459	0.0596	0.0491	0.0793	0.1197	0.0325	0.0428	0.0554	0.0754	0.1212	0.1843	0.0501	0.0659	0.0857
BERT4Rec	0.0558	0.0902	0.1372	0.0370	0.0490	0.0636	0.0519	0.0836	0.1261	0.0344	0.0453	0.0586	0.0802	0.1284	0.1947	0.0533	0.0700	0.0908
MacridVAE	0.0487	0.0796	0.1214	0.0321	0.0427	0.0556	0.0452	0.0738	0.1121	0.0299	0.0397	0.0515	0.0713	0.1148	0.1743	0.0473	0.0624	0.0811
CMVCDR	0.0601	0.0967	0.1466	0.0397	0.0523	0.0679	0.0563	0.0904	0.1367	0.0372	0.0490	0.0635	0.0861	0.1374	0.2075	0.0573	0.0750	0.0974

Table 3: Main recommendation results on three benchmarks (full-ranking protocol). Best results are bolded ; second best is underlined .

\dagger

: statistically significant improvement over the best baseline (

p<0.05

Variant	HR@10	$\Delta$ HR	NDCG@10	$\Delta$ NDCG
MOSAIC (full)	0.1343	—	0.0696	—
w/o Spe	0.1251	$-6.9\%$	0.0639	$-8.2\%$
w/o Com	0.1274	$-5.1\%$	0.0651	$-6.5\%$
w/o Cross	0.1187	$-11.6\%$	0.0601	$-13.6\%$
w/o GRL	0.1292	$-3.8\%$	0.0665	$-4.5\%$
w/o Align	0.1306	$-2.8\%$	0.0674	$-3.2\%$

Table 4: Ablation study on Amazon Movie–Book (Movie domain).

\Delta

denotes the relative drop from the full MOSAIC model.

Hyperparameter	Values tested	NDCG@10 range
$\beta_{1}$	0.001, 0.01, 0.1 , 1.0	0.0641 – 0.0696
$\beta_{2}$	0.001, 0.01 , 0.1, 1.0	0.0658 – 0.0696
$\beta_{3}$	0.01, 0.1 , 0.5, 1.0	0.0663 – 0.0696
$\beta_{4}$	0.01, 0.1 , 0.5, 1.0	0.0657 – 0.0696
$\rho$	0.1, 0.5, 1.0 , 2.0, 5.0	0.0638 – 0.0696
$d$	64, 128 , 256	0.0651 – 0.0701

Table 5: Hyperparameter sensitivity: NDCG@10 on Amazon Movie–Book (Movie) validation split under individual parameter variation, holding all others fixed at optimal values.

Recommender Systems

Recommender systems predict and rank items for people using interaction histories, content, context, and feedback.

See all 514 papers

Jun 1, 2026

Breaking the Information Silo: Semantic Personas for Cross-Domain Recommendation

Digital platforms increasingly operate as isolated information silos, limiting their ability to construct comprehensive user representations across domains. Cross-domain recommender systems seek to overcome this limitation by transferring knowledge from a source domain to a target domain, yet most existing approaches depend on shared users, shared items, or structurally similar interaction graphs. These assumptions are often unrealistic across independent platforms. We propose SPHERE (Semantic Personas for Heterogeneous cross-domain Recommendation), a design artifact that enables recommendation knowledge transfer across strictly disjoint domains with no shared users or items. Rather than aligning domains through identity or graph structure, SPHERE uses large language models to induce a shared behavioral vocabulary, generate structured semantic personas for users, and retrieve behaviorally similar source-domain communities that form a Community Source Persona. This semantic signal is integrated with collaborative signals through a dual-tower architecture and dynamic fusion gate, allowing SPHERE to augment standard recommender backbones. Empirical evaluation across Amazon Books, Goodreads, and Steam demonstrates consistent improvements over NCF, SVD++, and LightGCN baselines under full-ranking evaluation. The results show that cross-domain transfer effectiveness is not determined solely by semantic proximity between domains; rather, it depends critically on the structural density and native predictive strength of the target domain. The study contributes to information systems research by reframing cross-domain personalization as behavior-based semantic alignment, offering a practical mechanism for overcoming information silos while preserving interpretability and modularity.

Jonathan Mayo, Moshe Unger, Konstantin Bauman

Technology and Information Management Department, Coller School of Management, Tel Aviv University, Tel Aviv 6997801, Israel · Management Information Systems Department, Fox School of Business, Temple University, Philadelphia, Pennsylvania 19122

Cross-Domain Recommendation Cross-Domain Transfer Learning Recommender Systems User Preference Modeling

Ask PDF

Aug 10, 2026

CoRCi: Cross-Reconstruction of Coherent Interests Modeling in Cross-Domain Sequential Recommendation

Cross-Domain Sequential Recommendation (CDSR) aims to alleviate data sparsity by transferring dynamic user interests across related domains. A key challenge lies in effectively bridging these domains. In single-domain modeling, models cannot distinguish between domain-specific and domain-invariant interests. Recent methods merge domain-specific sequences chronologically into a mixed-domain sequence to capture domain-invariant knowledge. However, they typically deploy separate encoders for the mixed-domain sequence and train them with per-domain loss aggregation. This workflow magnifies inter-domain discrepancies and disrupts domain-invariant interest coherence, especially when query target pairs in Seq2Seq originate from different domains. In this paper, we present CoRCi (Cross-Reconstruction for Coherent Interest), a dual-target CDSR framework that tackles these drawbacks. Specifically, CoRCi proposes a Cross-Reconstruction approach that generates mixed-domain representations directly from pre-encoded specific-domain representations via cross-attention. The generated representations are then trained using a single, sequence-level, domain-agnostic loss to preserve the coherence of domain-invariant interests. To further suppress domain discrepancies in mixed-domain modeling, CoRCi introduces FocalNCE, which embeds Focal Loss into the preceding mixed-domain InfoNCE objective. The new loss assigns higher penalties to negatives drawn from the same domain as the query, thereby strengthening domain-invariant alignment. Extensive experiments on four real-world datasets demonstrate that CoRCi consistently outperforms state-of-the-art CDSR counterparts, achieving statistically significant gains across all metrics.

Qingtian Bian, Tieying Li, Marcus de Carvalho +3

Nanyang Technological University Singapore · Northeastern University China · Shanghai University of Finance and Economics China

Sequential Recommendation Invariant Representation Learning

Ask PDF

Oct 12, 2026

OT-FlowRec: Optimal Transport-Guided Multi-Bridge Flow for Multimodal Sequential Recommendation

E-commerce user preferences evolve dynamically within sessions, exhibiting concurrent and multi-centric interest distributions. While existing multimodal sequential recommendation methods leverage global attention to compress historical interactions, they inadequately capture the concurrent nature of multi-interest behaviors and fail to model the underlying multi-centric distribution of user intent. To tackle this issue, we propose OT-FlowRec, an optimal transport-guided multi-bridge flow matching framework that explicitly models multi-interest distributions to facilitate multimodal sequential recommendation. Unlike single-path generative models, OT-FlowRec constructs a mixture of parallel, interest-specific transport flows to model the multi-centric distribution characteristics, where each flow preserves a distinct user interest cluster via an optimal transport plan. Then, a geometric transport mechanism is designed to establish minimum-cost paths from the latent multi-interest distribution to target preference prototypes to provide multi-interest awareness. Moreover, a progressive time-sampling warmup strategy dynamically biases optimization weights toward the final states of the generation process, bridging the gap between continuous global distribution fitting and practical top-

k

ranking metrics. This approach achieves explicit interest disentanglement, ensuring the preservation of a multi-peaked preference representation instead of collapsing it into a unimodal average. Extensive experiments on three public datasets demonstrate that OT-FlowRec outperforms existing state-of-the-art baselines.

Miao Sun

Nanjing University of Science and Technology School of Computer Science and Engineering

Recommender Systems Multimodal Recommendation Flow-Based Generative Modeling Flow Matching Optimal Transport

Ask PDF

Abstract

Recommender Systems

Explore similar work

Breaking the Information Silo: Semantic Personas for Cross-Domain Recommendation

CoRCi: Cross-Reconstruction of Coherent Interests Modeling in Cross-Domain Sequential Recommendation

OT-FlowRec: Optimal Transport-Guided Multi-Bridge Flow for Multimodal Sequential Recommendation