Welcome to the Efficient AI Workshop!

The rapid growth of AI, driven by large datasets and foundation models, has brought remarkable advances but also steep computational, memory, and energy costs. This workshop focuses on efficient AI through data and model reduction, covering techniques such as pruning, quantization, distillation, parameter-efficient fine-tuning, dataset condensation, and coreset selection. By bringing together researchers from machine learning, data mining, and applied AI, we aim to foster discussion on scalable and data-efficient solutions for modern AI systems, including emerging challenges in foundation models and LLMs.

Call for Papers

We invite original, unpublished research papers on the challenges and advances in data and model efficiency for scalable AI. Accepted papers will be presented as posters, with a subset selected for oral presentations, and published in the IEEE BigData 2026 Workshop Proceedings by the IEEE Computer Society Press. All submissions are reviewed through a single-blind process by members of the program committee.

Topics of Interest

We encourage the submission of papers on topics within the areas listed below. Submissions are not restricted to these specific topics, and we welcome any work that aligns with the general goals of the workshop.

Model Efficiency and Compression

Model pruning, sparsification, and quantization
Knowledge distillation and model compression
Parameter-efficient fine-tuning (e.g., adapters, LoRA)
Low-rank and tensor-based model reduction
Efficient inference and dynamic computation (e.g., token pruning)

Data Efficiency and Reduction

Dataset condensation and distillation
Coresets and subset selection
Graph reduction (e.g., coarsening, sparsification, or condensation)
Efficient learning over graph-structured and relational data
Data pruning and synthetic data for efficient training

Efficiency in Foundation Models and LLMs

Efficient post-training, fine-tuning, and adaptation of foundation models
Prompt compression and context reduction
Retrieval-efficient RAG systems
Memory and KV-cache optimization
Efficient reasoning, retrieval, and agentic workflows for LLMs

Systems and Scalability for Efficient AI

Scalable pipelines for data and model efficiency
Distributed and memory-efficient training
Hardware-aware optimization

Applications of Efficient AI

Efficient large-scale real-world applications (e.g., remote sensing, healthcare, recommender systems)
Edge, mobile, and on-device learning
Resource-constrained AI for low-compute or low-connectivity environments
Practical deployments demonstrating efficiency, scalability, and usability

Trustworthy, Interpretable, and Measurable Efficient AI

Explainability of compressed or reduced models
Trade-offs between efficiency, accuracy, robustness, and cost
Benchmarks, metrics, and protocols for efficiency-quality-cost trade-offs
Theoretical insights into data and model reduction

Submission Guidelines

Format: Submissions must be a PDF file in IEEE 2-column format, following the IEEE Computer Society Proceedings Manuscript Formatting Guidelines (see formatting instructions).
Paper types: We accept both full papers (up to 10 pages) and short papers (up to 5 pages). References are counted within the page limit.
Review: This workshop uses single-blind review. Author names and affiliations must be included in the submission.
Presentation: At least one author of each accepted paper must register and present the work in person at Phoenix, AZ, USA.