4th Workshop on Handling Resource Constraints for/using Big Data and AI

Overview

The rapid evolution of Artificial Intelligence (AI) and Big Data technologies has significantly enhanced human productivity, accessibility, and decision-making across domains. However, this progress is accompanied by growing disparities due to limitations in computational resources, energy consumption, data availability, infrastructure, and security risks. These constraints increasingly impact the equitable deployment and accessibility of AI systems, particularly in low-resource and emerging environments.

The rise of foundation models, large language models (LLMs), multimodal systems, and edge AI has further amplified the need to design systems that are not only powerful but also efficient, adaptive, and sustainable. The challenge is no longer just building intelligent systems, but ensuring that they can operate under constraints such as limited memory, bandwidth, compute, labeled data, and energy budgets, while maintaining fairness, reliability, and robustness.

Advancing "technology for humanity as a whole" requires a multi-disciplinary approach, integrating innovations in efficient computing architectures, green and energy-aware AI, data-centric AI and low-resource learning, edge and distributed agentic systems, privacy-preserving and secure AI, and sustainable infrastructure and environmental considerations.

Workshop Objectives

The proposed CONSTRAINT-2026 workshop aims to:

• Provide a platform for showcasing cutting-edge research in AI and Big Data under resource constraints
• Foster interdisciplinary collaboration across academia, industry, and policy domains
• Encourage development of benchmarks, datasets, algorithms, and systems tailored to constrained agentic environments
• Promote innovations that bridge the gap between state-of-the-art AI and real-world deployability, especially in underserved regions

The broader objectives of CONSTRAINT-2026 are:

• To investigate systemic, computational, and societal constraints that limit the development and deployment of AI/Big Data systems
• To promote equitable, inclusive, and accessible AI systems, especially for low-resource and multilingual settings
• To enable collaboration across communities working on efficient, sustainable, and constraint-aware AI systems
• To advance a multi-disciplinary framework for deploying AI systems responsibly in real-world environments
• To encourage research on efficient foundation models and next-generation AI systems under constraints

Call for Papers

CONSTRAINT-2026 welcomes submissions on various topics contributing to developing Big Data and AI systems under resource constraints. We particularly encourage studies that address practical applications or improve upon resource constraints for AI/Big Data systems including, but not limited to:

Efficient Model Design & Scaling

• Scaling foundation models and LLMs under compute, memory, and latency constraints
• Model compression: pruning, quantization, distillation, token pruning, low-rank adaptation (LoRA)
• Sparse architectures, Mixture-of-Experts (MoE), adaptive computation
• Efficient training paradigms: few-shot, zero-shot, continual learning, test-time adaptation

Edge & Distributed AI

• AI deployment on edge devices, IoT systems, and mobile platforms
• Federated learning and decentralized training under communication constraints
• On-device inference, streaming AI, and real-time systems
• Resource-aware scheduling and orchestration

Green & Sustainable AI

• Energy-efficient training and inference strategies
• Carbon-aware AI systems and benchmarking
• Sustainable data centers and hardware-software co-design
• Lifecycle analysis of AI systems (training → deployment → recycling)

Data-Centric & Low-Resource Learning

• Learning under limited, noisy, or biased data
• Synthetic data generation and augmentation
• Active learning, semi-supervised and self-supervised learning
• AI for low-resource languages and domains

Multimodal & Generative AI Under Constraints

• Efficient multimodal models (vision-language, speech-language, sensor fusion)
• Retrieval-augmented generation (RAG) under resource limits
• Memory-efficient context handling and long-context modeling
• Lightweight reasoning and planning in large models

Privacy, Fairness & Robustness

• Privacy-preserving learning (differential privacy, secure computation)
• Robustness to adversarial and distributional shifts
• Fairness and bias mitigation under constrained settings
• Explainability and auditing in low-resource environments

Hardware, Systems & Benchmarking

• Specialized hardware (AI accelerators, neuromorphic computing)
• VLSI and architecture design for efficient AI workloads
• Compiler optimizations and runtime systems
• Storage and database optimization using AI
• Benchmarking efficiency, fairness, and sustainability
• New metrics for resource-aware performance evaluation
• Cost-performance trade-offs and Pareto optimization

Real-World Applications & Emerging Topics

• Real-world deployment case studies
• AI under regulatory, ethical, and governance constraints
• Human-AI collaboration under limited resources
• Continual, lifelong, and adaptive AI systems
• Neuro-symbolic and hybrid AI under constraints
• AI for scientific discovery with limited compute/data

Important Dates

• October 1, 2026: Workshop Paper Submission Due Date
• November 4, 2026: Notification of Acceptance
• November 25, 2026: Camera-ready Papers Due
• December 14–17, 2026 (Online) Workshop Dates

Invited Keynote Speakers

Program Chairs

• Manikandan Ravikiran, Research Scientist, Thoughtworks AI Labs, India
• Ankit Sharma, Senior Researcher, Hitachi Ltd, Japan
• Rakesh Prakash, Architect, Boeing, USA
• Sathyanarayanan Aakur, Associate Professor, Auburn University, USA
• Rohit Saluja, Assistant Professor, IIT Mandi, India

Program Committee (Preliminary)

• Rajat Verma, IIT Mandi, India
• Manoj Balaji, Juniper Networks, India
• Sheetal Kumar, Hitachi Ltd
• Yuta Koreeda, Hitachi Ltd, Japan
• Sena Ekiz, Nvidia, USA
• Anna Muyan Li, University of Waterloo
• Tarun Sharma, IIT Mandi, India
• Radhika Grover, IIT Mandi, India
• Ankit Maurya, HCL Technologies