IAAI-26 Program

Collocated with AAAI-26 | Singapore EXPO | Singapore

The Thirty-Eighth Annual Conference on Innovative Applications of Artificial Intelligence (IAAI-26) is a venue for papers describing highly innovative realizations of AI technology. The objective of the conference is to showcase successful applications and novel uses of AI. The conference will use technical papers, best practice papers, invited talks, and panel discussions to explore issues, methods, and lessons learned in the development and deployment of AI applications; and to promote an interchange of ideas between basic and applied AI and the discourse on the actual deployment of AI in practice. The general goal of the conference is to teach people the challenges and solutions to accomplishing something useful in the real world, as opposed to describing a new algorithm.

Resources

• IAAI-26 Program Schedule
• IAAI-26 Track Poster Presentation Schedule

IAAI Invited Talk Program

Thursday, January 22

Milind Tambe

Generative AI for Social Impact: Towards Solving the Deployment Bottleneck

3:05 – 4:00 PM

My team’s work on AI for Social Impact (AI4SI) has spanned two decades, focusing on optimizing limited resources in critical areas like public health, conservation, and public safety. I will present field results from India, where deployed restless and collaborative bandit algorithms achieved significant improvements in the world’s two largest mobile maternal health programs.

I will also present ongoing work on network-based HIV prevention in South Africa, modeled as a branching bandit problem. These projects, and other work across Africa and Asia that I will highlight, expose a critical bottleneck to AI4SI scaling—the deployment bottleneck—which spans all stages of the ML pipeline: the observational scarcity gap (data), the policy synthesis gap (learning/modeling), and the human-AI alignment gap (deployment).

This talk investigates how Generative AI can accelerate the AI4SI deployment cycle, specifically through the leveraging of LLM Agents and diffusion models. LLM Agents address the alignment gap by integrating expert guidance into algorithmic planning, yielding resource optimization strategies that reflect real-world priorities. Furthermore, diffusion models address the observational scarcity and policy synthesis gaps by generating synthetic social networks, applying Transfer RL to utilize data across domains, and efficiently synthesizing complex policies. I will conclude by discussing this path toward scalable and human-aligned AI for Social Impact.

Bio

Milind Tambe is Gordon McKay Professor of Computer Science at Harvard University; concurrently, he is also Principal Scientist and Director for “AI for Social Good” at Google Research. Prof. Tambe and his team have developed innovative AI and multi-agent reasoning systems that have been successfully deployed to deliver real-world impact in public health (e.g., maternal and child health), public safety, and wildlife conservation. He is the recipient of the AAAI Award for Artificial Intelligence for the Benefit of Humanity, the AAAI Feigenbaum Prize, the IJCAI John McCarthy Award, the AAAI Robert S. Engelmore Memorial Lecture Award, the AAMAS ACM/SIGAI Autonomous Agents Research Award, and the INFORMS Wagner Prize for excellence in Operations Research practice. He is a fellow of AAAI and ACM. For his work on AI and public safety, he has also received the Military Operations Research Society Rist Prize for best implemented national security operations research study, the Columbus Fellowship Foundation Homeland security award, and commendations and certificates of appreciation from the US Coast Guard, the Federal Air Marshals Service, and airport police at the city of Los Angeles.

Marko Grobelnik

AI Incidents: Analysis and Trends

5:35 – 6:30 PM

Abstract

The rapid evolution of Artificial Intelligence (AI) inevitably introduces new risks and challenges. As part of the initiative at the OECD AI Policy Observatory (oecd.ai), we have developed a comprehensive system to monitor AI-related incidents and hazards reported in global mainstream media on a daily basis. The system, publicly available at https://oecd.ai/incidents, is widely utilized by the international policy and research community. This presentation details the OECD’s multi-level framework for analyzing AI incidents, covering the formal OECD definitions of incidents and hazards, the methodology for data collection, and the underlying system architecture. Furthermore, we provide a detailed analysis of the collected data trends. A key finding reveals that while the absolute volume of reported AI incidents is increasing, the relative incidence rate (normalized against the total volume of detected AI-related events) remains stable.

Bio

Marko Grobelnik is a researcher in the field of Artificial Intelligence (AI). Focused areas of expertise are Machine Learning, Data/Text/Web Mining, Network Analysis, Semantic Technologies, Deep Text Understanding, and Data Visualization. Marko co-leads Artificial Intelligence Lab at Jozef Stefan Institute, cofounded UNESCO International Research Center on AI (IRCAI), and is the CEO of Quintelligence.com specialized in solving complex AI tasks for the commercial world. He collaborates with major European academic institutions and major industries such as Bloomberg, British Telecom, European Commission, Microsoft Research, New York Times, OECD. Marko is co-author of several books, co-founder of several start-ups and is/was involved into over 100 EU funded research projects in various fields of Artificial Intelligence. Significant organisational activities include Marko being general chair of LREC2016 and TheWebConf2021 conferences. Marko represents Slovenia in OECD AI Committee (AIGO/ONEAI), in Council of Europe Committee on AI (CAHAI/CAI), NATO (DARB), and Global Partnership on AI (GPAI). In 2016 Marko became Digital Champion of Slovenia at European Commission.

Friday, January 23

Panel on LLM Localization: Challenges and Opportunities

8:30 – 9:25 AM

This panel will explore the challenges, approaches, and opportunities of localizing LLMs. Discussion will initially focus on solving data, alignment, and other problems from a practical perspective before opening questions to the audience.

William Tjhi

Dr. William Tjhi, Head of Applied Research at AI Singapore, is pioneering the development of language-specific Large Language Models (LLMs) tailored for Southeast Asia’s diverse linguistic landscape. With over 15 years of experience in AI and machine learning at leading organizations like GovTech and Traveloka, he is dedicated to making AI accessible and inclusive for all.

Wenxuan Zhang

Wenxuan Zhang is currently a tenure-track SUTD Assistant Professor (SAP) at the Information Systems Technology and Design (ISTD) Pillar, Singapore University of Technology and Design (SUTD). He received his PhD degree from the Chinese University of Hong Kong, and then joined Alibaba Singapore as a research scientist with the prestigious Ali Star award. His primary research areas are natural language processing (NLP) and large language models (LLMs). His research aims to advance NLP models that are inclusive, supporting diverse languages and cultures through multilingual language models; while also trustworthy by improving the understanding, safety, and robustness of the models. He (co-)led multiple influential open-source research projects, including SeaLLMs (Large Language Models specialized for Southeast Asian languages), Babel, and AutoArena. He regularly serves as an area chair and on program committees for multiple leading conferences and journals, including ACL, EMNLP, NeurIPS, ICLR etc. He served on the organizing committee of SSNLP 2025 and organized tutorials at IJCAI 2023 and SIGIR 2025. He is also recognized among the World’s Top 2% Scientists (by Stanford and Elsevier) in 2025.

Longxu Dou

Longxu Dou is a Research Scientist at Sea AI Lab (Singapore) specializing in Large Language Models and Code Agents. He has (co-)led several high-impact open-source research projects, including Sailor (500K+ downloads), a suite of LLMs optimized for Southeast Asian languages, and Reptile, a terminal agent framework that integrates human-in-the-loop learning with reinforcement learning. Longxu completed his Ph.D. and B.S. at Harbin Institute of Technology and has held research internships at both Microsoft Research Asia and the National University of Singapore. Furthermore, he has published over 30 papers in top-tier AI venues, such as ICLR, NeurIPS, ICML, and ACL.

Pang Wei Koh

Pang Wei Koh is an assistant professor in the Allen School of Computer Science and Engineering at the University of Washington, a research scientist at the Allen Institute for AI, and a Singapore AI Visiting Professor. His research interests are in the theory and practice of building reliable machine learning systems. His research has been published in Nature and Cell, featured in The New York Times and The Washington Post, and recognized by the AI2050 Early Career Fellowship, MIT Tech Review Innovators Under 35 Asia Pacific award, Google ML and Systems Junior Faculty Award, and best paper awards at ICML, KDD, and ACL. He received his PhD and BS in Computer Science from Stanford University. Prior to his PhD, he was the 3rd employee and Director of Partnerships at Coursera.

Sean McGregor (Moderator, IAAI General Chair)

Daniel Borrajo

Understanding and Reasoning

5:35 – 6:30 PM

Large Language Models (LLMs) have revolutionized the world of AI applications due to their ability to understand, process, and generate language. This has enabled the automation of tasks that were recently only partially solved by small-scale techniques. The ability to manage natural language has led to the belief that they are also capable of reasoning. As a result, numerous AI teams are using LLMs as general problem solving tools that solve all kinds of tasks. In this talk, a discussion will be presented on whether LLMs can (and/or should) be used to solve tasks other than those related to language. Hybrid models will be proposed, showing the advantages of classical AI techniques and LLMs when integrated into a single solution. In particular, the finance domain regularly deals with risks, and thus it requires AI reasoning to provide validity guarantees in many cases. The integration of LLMs with classical AI techniques could offer enhanced solutions that address the unique challenges faced by the financial industry, ensuring more reliable and risk-averse outcomes.

Bio

Daniel Borrajo is a Managing Director at JPMorganChase AI Research. He has been working at JPMorganChase since 2019 when he joined the AI Research team. He received his PhD in Computer Science in 1990 from Universidad Politecnica de Madrid and was a Professor at Universidad Carlos III de Madrid (UC3M). He has more than 40 years of research experience on AI. His main research interests are in the integration of the two main AI paradigms: model-based (e.g. AI Planning, Rules) and model-free (e.g. Machine learning). He has published over 300 papers. He has been Program Chair of AI-related international conferences (ICAPS), member of the ICAPS Council, regularly serves in the program committee of leading international AI conferences (IJCAI, AAAI, ICAPS, …), and he was Associate Editor of the Artificial Intelligence Journal until 2024.

Saturday, January 24

Bryan Goodman

Agentic AI at Enterprise Scale: Three Production Cases from Automotive

8:30 – 9:25 AM

This talk presents three production deployments of agentic AI at Ford Motor Company. It will focus on practical architectures, safeguards, and measurable impact. We define agentic AI as LLM driven systems with tool use, planning, and memory, orchestrated to perform multi step tasks reliably at scale. First, CodeGuardians is an AI powered code review accelerator. It combines static analysis, policy checks, and LLM critique in a reviewer in the loop workflow to improve development velocity and code quality. Second, FordAI, a customer assistant embedded on Ford.com and other applications, unifies the use of tool calls to deliver personalized support. It operates under strong safety and privacy guardrails, increasing self-service and customer satisfaction. Third, an AI enabled engineering requirements platform integrates LLM reasoning with knowledge and constraint checking to detect conflicting software feature requirements and enable deep root cause analysis. This shortens diagnosis and resolution cycles. We distill lessons on agent orchestration, evaluation, safety and governance, observability, and cost/latency optimization. The talk will offer practical guidance for researchers and practitioners on deploying agentic AI in production at enterprise scale.

Bio

Bryan Goodman is Executive Director of Artificial Intelligence at Ford Motor Company, where he leads the enterprise AI strategy, oversees Ford’s AI ethics policy as chair of the AI Technology and Ethics Council, and directs teams delivering scalable, high‑impact AI solutions across the business. He is responsible for Ford’s AI and machine learning platform and for providing generative AI capabilities company‑wide, including multimodal systems for audio, image, video, and 3D. Goodman drives incorporation of AI into products and services spanning engineering and design, marketing and forecasting, customer assistance (web, contact centers, apps, in‑vehicle), and manufacturing.

Since joining Ford in 1999, Goodman has built and scaled enterprise AI capabilities, founded the AI Advancement Center, and set Ford’s first enterprise AI and formal data strategies approved by the C‑suite and Board. An inventor on 13 U.S. patents and recipient of the Henry Ford Technology Award, he holds a PhD in Physical Chemistry and Computational Science & Engineering from the University of Illinois and a BS in Mathematics and Chemistry, summa cum laude, from Hope College.

David Atienza

Toward Personalized and Sustainable Healthcare with Edge AI Medical Wearables

9:30 – 10:25 AM

The convergence of edge AI and smart medical devices is revolutionizing personalized healthcare by enabling real-time, energy-efficient data processing directly on wearable platforms. This keynote explores two complementary architectural strategies that are accelerating innovation in edge AI systems for medical applications.

The first strategy enhances general-purpose computing through in-memory computing, bringing intelligence closer to the data source. By embedding computation within memory blocks and leveraging validated open-hardware microarchitectures, this approach enables low-latency and energy-aware processing, which are critical features for continuous health monitoring and real-time decision-making.

The second strategy focuses on full-system co-design, integrating analog and digital components to develop domain-specific, heterogeneous System-on-Chip (SoC) platforms. These platforms incorporate specialized accelerators as co-processors, optimized for biomedical signal processing, sensor fusion, and other healthcare-specific tasks. This enables the creation of robust, adaptable, and efficient wearable solutions.

Building on recent advances in open-hardware SoC frameworks, the keynote demonstrates how the aforementioned two strategies can be synergistically combined to develop the next-generation of medical-grade wearables capable of running advanced deep learning algorithms. The talk concludes with a forward-looking perspective on how open-source hardware and accelerator-centric edge AI can empower distributed and federated learning, paving the way for more personalized, privacy-preserving healthcare. Applications span chronic disease management, preventive diagnostics, and real-time monitoring in cardiovascular, neurological, and musculoskeletal disorders.

Bio

David Atienza is a professor of Electrical and Computer Engineering, heads the Embedded Systems Laboratory (ESL) and is the Associate Vice President of Research Centers and Platforms for the period 2024-2028 at EPFL, Switzerland. His research interests include system-level design methodologies for multi-processor system-on-chip (MPSoC) targeting low-power Cyber-Physical Systems (CPS) and energy-efficient computing servers. His latest works include new 2.5D/3D power/thermal-aware design and architectures for MPSoCs targeting edge AI systems, as well as HW/SW co-design and AI-based multi-level optimization for sustainable computing in the Internet of Things (IoT) context.

Prof. David Atienza has co-authored over 450 papers, one book, and 14 patents in these previous areas. He has also received multiple recognitions and awards, among them the IEEE/ACM HW/SW Co-Design Conference (CODES-ISSS) 2024 Test-of-Time Award for the most influential paper in the last 15 years, the ICCAD 10-Year Retrospective Most Influential Paper Award in 2020, the Design Automation Conference (DAC) Under-40 Innovators Award in 2018, and IEEE CEDA and ACM SIGDA Early Career Awards on EDA tools and systems research. He is currently the Editor-in-Chief of IEEE Trans. on CAD (T-CAD) and ACM Computing Surveys. He is a Fellow of IEEE, a Fellow of ACM, and was the Chair of the European Design Automation Association (EDAA) from 2022 until 2024, and the President of IEEE CEDA from 2017 to 2018. Finally, he has created two successful EPFL spin-offs that commercialize medical wearables: SmartCardia, which commercializes cardiovascular wearables for home-based monitoring, and Sensemodi, which is a medical start-up that provides personalized monitoring using multi-sensing wearable technologies for knee chondropathy.

Ece Kamar

Navigating the AI Horizon: Promises, Perils, and the Power of Collaboration
2:00 – 2:55 PM

We stand at the dawn of the AI era, a technological revolution poised to be the most consequential of our generation, presenting both unprecedented opportunities and profound challenges. But this promise is shadowed by significant challenges. To build a future we want, we must move beyond the hype and the headlines to confront the most pressing open problems—technical, sociotechnical, and multidisciplinary. This talk will review the rapid progress, dissect challenges ahead, and argue that our greatest task isn’t simply building smarter machines, but fostering the human wisdom to guide them towards a future that is not only intelligent but also equitable, safe, and profoundly human.

Bio

Ece Kamar is Distinguished Scientist, Corporate Vice President and the Managing Director of the AI Frontiers Lab at Microsoft Research. She leads research and development towards pushing the frontiers of AI capabilities. Releases from her lab includes Phi family of models and libraries to empower agentic work, including AutoGen, MagenticOne and MagenticUI. Ece has a decade of experience studying the impact of AI on society and developing AI systems that are reliable, unbiased and trustworthy. She has been instrumental in building the Responsible AI efforts inside Microsoft. She serves as Technical Advisor for Microsoft’s Internal Committee on AI, Engineering and Ethics. Ece received a Bachelor of Science in Computer Engineering from Sabanci University in Istanbul. She completed her PhD at Harvard University, where she has worked with Prof. Barbara Grosz on human-AI collaboration. Ece is an Affiliate Faculty in the Department of Computer Science and Engineering at the University of Washington and is an Adjunct Faculty at Sabanci University. She was a member of the first study panel of Stanford’s AI100. She is currently serving on the National Academies’ Computer Science and Telecommunications Board (CSTB). She has been on the organizing committees of top AI and HCI conferences and received multiple best paper awards for her work.

Robert Englemore Memorial Lecture Award

Ashok Goel

AI for Reskilling, Upskilling, and Workforce Development

As AI becomes increasingly powerful and ubiquitous, it is disrupting skills and displacing workers. NSF’s National AI Institute for Adult Learning and Online Education (AI-ALOE) posits that AI can be part of the solution to the growing problem if we can use AI for reskilling, upskilling, and workforce development at scale. The long-term vision of AI-ALOE is to develop and use AI technologies to enhance the proficiency of online education for all adult learners, using in-person education as a benchmark. The day-to-day mission of AI-ALOE is to conduct responsible research into AI that is grounded in theories of human cognition and learning and derived from the scientific process of learning engineering. I will describe ongoing research at AI-ALOE.

Bio

Ashok Goel is a Professor of Computer Science and Human-Centered Computing in the School of Interactive Computing at Georgia Institute of Technology, and the Chief Scientist with Georgia Tech’s Center for 21st Century Universities. He is a Fellow of AAAI and the Cognitive Science Society, an Editor Emeritus of AAAI’s AI Magazine, and a recipient of multiple IBM Faculty Awards and AAAI’s Outstanding AI Educator Award as well as Distinguished Service Award. Ashok is the PI and Executive Director of NSF’s National AI Institute for Adult Learning and Online Education (aialoe.org) headquartered at Georgia Tech. He is the Founder of Beyond Question AI, LLC.