Call for Papers | Download the CFP Flyer

ICMAI 2027 welcomes original papers reporting high-quality, previously unpublished research on mathematics and artificial intelligence. The conference invites submissions across the following nine tracks. Topics of interest include, but are not limited to:

Track 1: Algorithms and Computational Complexity

This track focuses on the design, analysis, and application of algorithms, with emphasis on computational complexity and optimization methods that underpin modern AI systems. Submissions are expected to make algorithmic or complexity-theoretic contributions, whether through new algorithm designs, tight complexity analyses, or optimization methods with provable guarantees. Topics of interest include (but are not limited to):

Design and analysis of exact, approximation, randomized, online, and streaming algorithms
Computational complexity: complexity classes, lower bounds, hardness of approximation, parameterized complexity
Combinatorial and discrete optimization; matroid and graph algorithms
Optimization for AI: convex and non-convex optimization, stochastic and distributed optimization, zeroth- and first-order methods
Quantum and quantum-inspired algorithms
Numerical linear algebra and fast computation for large-scale learning
Game theory and algorithmic mechanism design

Track 2: Logic, Reasoning and Knowledge Representation

This track welcomes work on applied logic and formal reasoning methods that enable reliable, interpretable, and verifiable AI. It covers logical foundations as well as practical reasoning systems, ranging from SAT/SMT solvers and theorem provers to knowledge graphs and neuro-symbolic integration. Topics of interest include (but are not limited to):

Applied and non-classical logics for AI (modal, temporal, description, fuzzy, and probabilistic logics)
Automated reasoning: SAT/SMT solving, theorem proving, model checking
Knowledge representation and reasoning
Knowledge graphs, ontologies, and semantic web technologies
Neuro-symbolic AI and logic-integrated learning
Formal specification and verification of AI systems
Commonsense, spatial, and temporal reasoning
Automated planning and causal reasoning

Track 3: Machine Learning and Deep Learning

This track covers models, methods, and theory of machine learning and deep learning, with priority given to methodological and applied contributions. It welcomes advances in learning architectures, training methods, and learning theory that are grounded in concrete models or applications. Topics of interest include (but are not limited to):

Supervised, unsupervised, and self-supervised learning
Deep learning architectures: Transformers, CNNs, GNNs, and diffusion models
Statistical learning theory: generalization bounds, PAC-Bayes analysis, and sample complexity
Optimization and training dynamics of learning models
Reinforcement learning and decision-making
Few-shot, meta-, transfer, and continual learning
Generative models and probabilistic methods
Machine learning with formal guarantees

Track 4: Computer Vision and Image Analysis

This track addresses the theory, algorithms, and applications of computer vision and image analysis. It covers the full pipeline from low-level image processing to high-level visual understanding, including multimodal perception and domain-specific vision systems. Topics of interest include (but are not limited to):

Image and video understanding: classification, detection, and segmentation
3D vision: reconstruction, depth estimation, point cloud processing, NeRF, and 3D generation
Medical and biological image analysis
Vision-language models and multimodal perception
Low-level vision: restoration, enhancement, and super-resolution
Visual tracking, pose estimation, and action recognition
Efficient inference and model compression for vision
Remote sensing and industrial vision applications

Track 5: Natural Language Processing and Speech

This track focuses on computational methods for understanding, generating, and reasoning with human language, in both text and speech. It welcomes work on large language models and their applications, as well as classical NLP and speech processing problems. Topics of interest include (but are not limited to):

Large language models: pretraining, fine-tuning, alignment, and evaluation
Machine translation, text generation, and summarization
Speech recognition, synthesis, and speaker processing
Question answering, information extraction, and retrieval-augmented generation
Multilingual and low-resource NLP
Dialogue systems and LLM-based agents
Reasoning, planning, and tool use in language models
Interpretability and safety of NLP systems

Track 6: AI for Science and Engineering

This track highlights AI methods that accelerate scientific discovery and engineering practice. It welcomes work where AI is applied to, or co-developed with, a scientific or engineering domain, including data-driven modeling, simulation, and domain-informed learning methods. Topics of interest include (but are not limited to):

AI for mathematics: conjecture discovery, theorem proving assistance, and symbolic computation
AI for physics, chemistry, and materials science
AI for biology, genomics, and drug discovery
AI for climate, environment, and earth sciences
Scientific machine learning: physics-informed neural networks and neural operators
Surrogate modeling, simulation, and digital twins
AI for industrial design, manufacturing, and engineering optimization

Track 7: Data Science and Big Data Analytics

This track covers methods and systems for extracting knowledge from large-scale and complex data. It spans the full data pipeline, from storage and processing frameworks to mining algorithms and analytics applications, with emphasis on scalability and real-world data challenges. Topics of interest include (but are not limited to):

Big data processing frameworks and distributed analytics
Data mining, pattern discovery, and knowledge discovery in databases
Time-series analysis, anomaly detection, and forecasting
Graph analytics and network science
Streaming and real-time analytics
Data quality, data integration, and data lakes
Scalable databases and query processing for AI workloads
Recommender systems and user behavior analytics

Track 8: Trustworthy, Responsible and Explainable AI

This track addresses methods and frameworks for making AI systems reliable, fair, transparent, and accountable. It brings together technical approaches to trustworthiness — including explainability, robustness, privacy, and formal verification — with broader questions of AI ethics, safety, and governance. Topics of interest include (but are not limited to):

Explainability and interpretability of AI models
Fairness, bias detection and mitigation, and AI ethics
Privacy-preserving machine learning: differential privacy and federated learning
Robustness, adversarial attacks and defenses, and out-of-distribution generalization
Formal verification, certification, and logic-based consistency checking of AI systems
AI safety, alignment, and governance
Uncertainty quantification and calibrated prediction
Accountability, auditability, and regulatory compliance of AI

Track 9: Interdisciplinary Applications

This track welcomes applied AI research that crosses disciplinary boundaries, addressing real-world problems in domains where AI methods intersect with domain knowledge, human factors, or societal needs. Contributions should demonstrate both methodological soundness and meaningful domain impact. Topics of interest include (but are not limited to):

AI in education and intelligent tutoring systems
AI in healthcare, medicine, and public health
AI in finance, economics, and business analytics
AI in agriculture, smart cities, and transportation
AI in law, social sciences, and the humanities
AI in art, design, and creative industries
AI for energy systems and sustainability
Robotics, embodied AI, and human-AI interaction