ICMLAA 2027 | Machine Learning and Autonomous Agents Conference|Tokyo, Japan

Conference Overview

The International Conference on Machine Learning and Autonomous Agents (ICMLAA 2027) will be held in Tokyo, Japan from July 23-25, 2027, themed on From Foundation Models to Autonomous Agents: Learning, Reasoning and Execution. Benefiting from the continuous innovation of core machine learning theories and algorithms, foundation models have greatly boosted the development of autonomous agent technology. Centered on core machine learning methodologies including representation learning, sequential decision learning and reasoning optimization, the conference focuses on how cutting-edge machine learning approaches empower the design, training and practical execution of various autonomous agents, tracking latest research progress and future trends in this interdisciplinary field.

This conference will gather top scholars, senior engineers, industry leaders and young students from more than 30 countries and regions. Through keynote speeches, special invited reports, oral paper presentations, poster sessions and industrial round‑table forums, participants will share the latest academic achievements, algorithm innovations and engineering practices covering core machine learning directions: reinforcement learning for agent decision, transfer & fine-tuning machine learning of foundation models, multi-agent collaborative machine learning algorithms, small/zero-shot machine learning for embodied agents, world modeling based on statistical machine learning, lightweight machine learning for edge agent deployment, robust and trustworthy machine learning for agent safety, together with practical implementation of machine-learning-powered industrial autonomous robots and digital human agents, to deepen industry‑university‑research cooperation and accelerate the transformation of academic achievements into practical industrial applications of machine learning-driven autonomous intelligence.

Download Call for Papers

Call For Papers

International Conference on Machine Learning and Autonomous Agents (ICMLAA 2027)
From Foundation Models to Autonomous Agents: Learning, Reasoning and Execution

Pre-training mechanisms and unsupervised learning of foundation models
Supervised fine-tuning and parameter-efficient tuning (PEFT)
Large model alignment, constraint learning and value optimization
Model distillation, pruning and sparse learning for foundation models
Transfer learning and domain adaptation of pre-trained models
Continual learning and incremental optimization for large models
Hyperparameter optimization and model generalization enhancement
Efficient training algorithms for low-resource foundation models

LLM-based agent architecture design and iterative upgrading
Autonomous tool calling and multi-function invocation learning
Agent memory mechanism, experience accumulation and retrieval learning
Task decomposition, automatic planning and autonomous execution
Long-horizon autonomous decision learning for intelligent agents
Dynamic goal adjustment and self-correction learning of agents
Generative agent behavior simulation and autonomous iteration
General-purpose agent construction based on foundation models

Multi-agent cooperative learning and group decision optimization
Multi-agent game learning, adversarial and collaborative equilibrium
Heterogeneous multi-agent fusion learning and task allocation
Decentralized and distributed multi-agent learning algorithms
Communication-aware multi-agent interactive learning
Scalable learning for large-scale multi-agent systems
Multi-agent meta-learning and fast adaptive collaboration
Conflict detection and coordination learning of multi-agents

Deep reinforcement learning for sequential autonomous decision-making
Model-based and model-free reinforcement learning optimization
Hierarchical reinforcement learning for complex task execution
Exploration-exploitation balance learning in autonomous systems
Reward function design and self-rewarding learning
Policy gradient optimization and robust policy learning
Imitation learning and inverse reinforcement learning for agents
Real-time reinforcement learning for dynamic autonomous scenarios

Embodied perception learning and environment sensing modeling
Robot motion control learning and autonomous manipulation
Vision-based embodied intelligent decision and interaction learning
Mobile robot autonomous navigation and path planning learning
Manipulator operation learning and precise autonomous control
Dynamic environment adaptive embodied learning
Human-robot interaction learning and collaborative operation
Embodied agent generalization learning in open environments

Few-shot learning for intelligent agent task adaptation
Zero-shot generalization learning of autonomous agents
Meta-learning and fast adaptive learning for new agent tasks
Data-efficient learning for low-resource agent scenarios
Cross-domain transfer learning for heterogeneous agent tasks
Weakly supervised learning for agent decision optimization
Semi-supervised learning for autonomous system iteration
Generalizable feature learning for unseen agent tasks

Neural world model construction and environment simulation learning
Future state prediction learning for dynamic environments
Physical consistency learning of simulated world models
Real-world environment reconstruction and predictive modeling
World model guided agent planning and decision learning
Temporal sequence learning and dynamic trend prediction
Multi-scale environment modeling and scenario reasoning learning
Simulation-to-real transfer learning for world model deployment

Cross-modal fusion learning of vision, text and sensor data
Multimodal feature alignment and unified representation learning
Multimodal prompt learning for autonomous agent understanding
Audio-visual fusion learning for environmental perception agents
Multimodal intent recognition and interactive learning
Cross-modal retrieval and matching learning for agent tasks
Multimodal generation learning for agent behavior output
Robust multimodal learning under noisy environment

Model lightweight compression and low-latency learning
Edge-side adaptive inference and incremental learning
Terminal autonomous learning with limited computing resources
Edge-cloud collaborative machine learning for agents
Low-power learning algorithms for embedded intelligent agents
Model quantization and binarization optimization learning
Real-time learning iteration for edge deployment scenarios
Resource-aware dynamic learning scheduling for agents

Robust learning against adversarial attacks on agent systems
Explainable machine learning for agent decision interpretability
Causal learning and causal reasoning for reliable agents
Privacy-preserving learning and federated learning for agents
Uncertainty estimation and risk-aware learning
Fault tolerance learning and abnormal behavior detection
Safety constraint learning for autonomous agent execution
Ethical and normative learning for intelligent agent behavior

Logical reasoning learning and deductive reasoning optimization
Inductive and abductive reasoning for complex agent tasks
Chain-of-thought learning and step-by-step reasoning iteration
Autonomous task planning and hierarchical logic learning
Knowledge graph enhanced agent reasoning learning
Long-term logical consistency learning for agent decisions
Reasoning error correction and self-verification learning
Cross-task reasoning generalization and migration learning

Machine learning for industrial autonomous robot control
Intelligent manufacturing agent learning and autonomous scheduling
Autonomous detection and fault diagnosis learning for industry
Smart logistics and warehouse autonomous agent applications
Industrial digital twin agent learning and interactive iteration
Autonomous inspection agent learning for industrial equipment
Smart city and traffic autonomous decision agent systems
Practical optimization and engineering deployment of industrial ML agents

Special Sessions

Proposal Template Download

LLM-Driven Embodied Intelligent Agent & Robot Learning

Large Model Empowered Embodied Intelligence and Robotic Learning

▶

• Large model empowered robot perception and motion decision
• Embodied agent sim-to-real transfer learning under complex environment
• Multimodal instruction-driven autonomous robotic manipulation
• Open-world generalization learning for physical intelligent agents
• Lightweight embodied model deployment on edge robot terminals
• Human-robot natural interactive learning based on foundation model
• World model construction for robotic long-horizon planning
• Industrial robot intelligent fault prediction with agent learning

Trustworthy Foundation Model & Secure Autonomous Multi-Agent System

Reliable Base Model and Secure Multi-Agent Collaborative System

▶

• Hallucination suppression and factual optimization of large foundation models
• Adversarial defense and robustness optimization for multi-agent cooperation
• Federated learning privacy protection for distributed intelligent agent clusters
• Causal inference enhanced reliable decision of autonomous agents
• Safety constraint alignment for LLM-based autonomous agents
• Interpretability research of multi-agent collaborative decision process
• Data security and copyright protection in model fine-tuning
• Risk early warning mechanism for open-ended agent autonomous execution

Conference Information

JOIN AS A PRESENTER

If you are interested in submitting a presentation proposal to ICMLAA 2027, you may choose to attend the conference as a Presenter. It's the best opportunity to share your novel ideas and best practices with a global audience.

We accept three types of presentations:

• Oral Presentations: 15 mins + 3 mins Q&A
• Poster Presentations: A1 size, 2 mins lightning talk
• Invited Talks: 30 mins + 5 mins Q&A

Please submit an abstract before registration. For inquiries, contact us at: contact_icmlla@yeah.net

CONFERENCE PROGRAM

July 23, 2027 | Friday

▶

09:00-17:00 Check-in & Materials
09:00-12:00 Pre-conference Tutorials
17:30-19:30 Welcome Reception

July 24, 2027 | Saturday

▶

08:30-09:00 Opening Ceremony
09:00-12:30 Keynotes & Invited Talks
14:00-17:30 Parallel Sessions
17:30-19:00 Poster Session
19:00-21:00 Award Ceremony

July 25, 2027 | Sunday

▶

09:00-12:30 Special Sessions & Roundtable
14:00-16:30 Technical Excursion
16:30-17:00 Closing Ceremony

Click the arrow to view detailed schedule

POSTER PRESENTATION

The conference poster must be designed in A1 size (841mm × 594mm) (portrait orientation) to meet the display requirements.

A1: 841mm × 594mm (Portrait)

Key Dates:

• Setup: July 24, 14:00-17:00
• Presentation: July 24, 17:30-19:00
• Removal: July 24, after 19:00

Outstanding posters will receive the "Best Poster Award" with certificates and prizes.

Download Poster Template

Conference Overview

Call For Papers

Track 1: Foundation Model Learning and Optimization

Track 2: Large Model Driven Autonomous Agents

Track 3: Machine Learning for Multi-Agent Collaboration

Track 4: Reinforcement Learning for Autonomous Decision-Making

Track 5: Embodied Machine Learning and Robotic Agents

Track 6: Small-Sample and Zero-Shot Machine Learning for Agents

Track 7: World Modeling and Predictive Learning for Autonomous Systems

Track 8: Multimodal Machine Learning for Intelligent Agents

Track 9: Lightweight and Edge Machine Learning for Agent Deployment

Track 10: Trustworthy, Robust and Safe Machine Learning for Agents

Track 11: Agent Reasoning, Planning and Logical Learning

Track 12: Machine Learning Based Industrial Autonomous Applications

Conference Proceedings

Special Sessions

LLM-Driven Embodied Intelligent Agent & Robot Learning

Trustworthy Foundation Model & Secure Autonomous Multi-Agent System

Important Dates

November 20th, 2026

December 20th, 2026

January 20th, 2027

Conference Information

JOIN AS A PRESENTER

CONFERENCE PROGRAM

POSTER PRESENTATION

🏆 Awards

👥 Join the Committee