New Hamiltonian Spectral Recommender models complex user behavior dynamics

Researchers have introduced the Hamiltonian Spectral Recommender (HSR), a novel approach to sequential recommendation that models user preference evolution using second-order dynamical systems. Unlike existing models that treat preference changes as first-order processes, HSR conceptualizes this evolution as a dissipative Hamiltonian system in a latent phase space, accounting for inertia, periodicity, and sudden shifts. The system's linear time-invariant structure allows for a closed-form solution in the frequency domain, while a learnable dissipation mechanism and an impulse refinement module capture interest decay and abrupt behavioral fluctuations. Experiments on benchmark datasets show HSR outperforms state-of-the-art Transformer and State Space Model-based recommenders. AI

IMPACT Introduces a novel modeling approach for sequential recommendation, potentially improving accuracy and capturing more complex user behavior patterns.

RANK_REASON Academic paper detailing a new model for sequential recommendation. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.IR (Information Retrieval) →

• alphaXiv
• arXiv
• CatalyzeX
• CORE Recommender
• DagsHub
• Gotit.pub
• Hamiltonian Spectral Recommender
• Hugging Face
• Influence Flower
• ScienceCast
• State Space Model
• Transformer++

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →