LT
ICDM 2026 Workshop
LLM for Tensor Analysis & Tensor for LLM Advancement
Call for Papers

ICDM 2026 Workshop on LLM for Tensor Analysis & Tensor for LLM Advancement

Overview

The LLMTenA 2026 workshop is motivated by the untapped synergy between tensor methods and large language models. While tensors offer mathematically grounded tools for dimensionality reduction, multi-modal fusion, and interpretability, LLMs provide powerful semantic understanding and generative capabilities.

We invite original submissions on topics including tensor-based LLM compression, LLM-enhanced tensor analysis, multi-modal fusion, interpretability, and applications across diverse domains such as recommendation systems and spatio-temporal data.

Topics of interest

  • Tensor-based compression and acceleration of LLMs
  • LLM-enhanced tensor decomposition, denoising, and completion
  • Multi-modal data fusion using tensors and LLMs
  • Interpretability and explainability of LLMs via tensor methods
  • Tensor networks for efficient fine-tuning and prompt engineering
  • Knowledge graph embedding and reasoning with tensors and LLMs
  • Applications in recommendation systems, neuroscience, spatio-temporal data, and multi-agent systems

Important dates

Anytime on Earth

MilestoneDate
Submission deadlineAugust 20, 2026
NotificationSeptember 20, 2026
Camera-ready deadlineOctober 05, 2026
Workshop dayNovember 12, 2026

Submission portal

Submit via CyberChair

Proceedings and presentation format

Accepted papers will be published in the joint ICDM workshop proceedings and archived in IEEE Xplore. All accepted papers will be divided into oral presentations and poster presentations, based on the reviewers' recommendations. In addition, the workshop will select and award one Best Oral Paper and one Best Poster Paper to recognize outstanding contributions. The awards will be announced during the closing session of the workshop.

Accepted Papers

The following five papers have been accepted to LLMTenA 2026. Please refer to the workshop program for scheduled oral presentations.

  1. TensorRecover: Recovering Tensorized Low-Rank Language Models
    Guanyi Fan, Tongkai Ji, and Wei Dai
  2. LLM-Elicited Constraint Priors for Reconstructing Physical Field Tensors
    Tianrui Chen, Siyuan Li, Panqi Chen, Yifan Sun, Mingjie Shao, and Lei Cheng
  3. The Retriever Should Remember: Experience-Amortized Reranking for Long-Term Agent Memory
    Qi Feng, Chris Ding, and Jicong Fan
  4. Co-Activation Predicts Functional Coherence: Pairwise Fusion for Compression in Mixture-of-Experts
    Evgeny Bessonnitsyn, Konstantin Pakulnevich, Marina Frolova, Ivan Shchukin, Timur Gabbasov, Aleksandr Ustinov, and Sergey Muravyov
  5. GLASS: Graph-Language Alignment with Spherical Scoring for Transferable Graph-Level Anomaly Detection
    Xudong Wang, Chris Ding, Tongxin Li, and Jicong Fan