EC1: Weinan Zhang

Shanghai Jiao Tong University

Large Decision Models
Over recent decades, sequential decision-making tasks are mostly tackled with expert systems and reinforcement learning. However, these methods are still incapable of being generalizable enough to solve new tasks at a low cost. In this article, we discuss a novel paradigm that leverages Transformer-based sequence models to tackle decision-making tasks, named large decision models. Starting from offline reinforcement learning scenarios, early attempts demonstrate that sequential modeling methods can be applied to train an effective policy given sufficient expert trajectories. When the sequence model goes large, its generalization ability over a variety of tasks and fast adaptation to new tasks has been observed, which is highly potential to enable the agent to achieve artificial general intelligence for sequential decision-making in the near future.

EC2: Haifeng Xu

University of Chicago

The Economics of Machine Learning
In this talk, I will overview our lab’s research on the economics of machine learning which consists of two complementary threads, namely, \emph{machine learning for economics} and \emph{economics for machine learning}. The first thread develops new learning algorithms for intelligent decision making in unknown strategic/economic environments ranging from foundational game-theoretic models, to recommender systems, and to defender-adversary interactions. I will highlight new techniques for addressing unique challenges in this pace such as non-stationary yet still traceable opponent behaviors and conflicting agent interests. The second thread of our lab’s research looks to employ economic principles to understand machine learning, such as the valuation and pricing of data. I will also briefly highlight our recent effort towards building a data trading platform as well as the interesting challenges we faced during this process.

EC3: Oren Salzman

Technion—Israel Institute of Technology

Algorithmic Motion Planning Meets Minimially-Invasive Robotic Surgery
Robots for minimally-invasive surgery such as steerable needles and concentric-tube robots have the potential to dramatically alter the way common medical procedures are performed. They can decrease patient-recovery time, speed healing and reduce scarring. However, manually controlling such devices is highly un-intuitive and automatic planning methods are in need. For the automation of such medical procedures to be clinically accepted, it is critical from a patient care, safety, and regulatory perspective to certify the correctness and effectiveness of the motion-planning algorithms involved in procedure automation. In this paper, I survey recent and ongoing work where we develop efficient and effective planning capabilities for medical robots that provide provable guarantees on various planner attributes as well as introduce new and exciting research opportunities in the field.

EC4: Lingjuan Lyu

Sony Research

A Pathway Towards Responsible AI Generated Content
AI Generated Content (AIGC) has received tremendous attention within the past few years, with content ranging from image, text, to audio, video, etc. Meanwhile, AIGC has become a double-edged sword and recently received much criticism regarding its responsible usage. In this article, we focus on three main concerns that may hinder the healthy development and deployment of AIGC in practice, including risks from privacy; bias, toxicity, misinformation; and intellectual property (IP). By documenting known and potential risks, as well as any possible misuse scenarios of AIGC, the aim is to sound the alarm of potential risks and misuse, help society to eliminate obstacles, and promote the more ethical and secure deployment of AIGC.

EC5: Panagiotis Kouvaros

Imperial College London

Towards Formal Verification of Neuro-symbolic Multi-agent Systems
Significant advances in Artificial Intelligence (AI) have enabled the automation of challenging tasks, such as computer vision, that have been traditionally difficult to tackle using classical approaches. This accelerated the trend of incorporating AI components in diverse applications with high societal impact, such as healthcare and transportation. Still, even thoughthere is an increasing consensus in AI being beneficial for society, its inherent fragility hinders its adoption in safety-critical applications. In response to these concerns the area of formal verification of AI has grown rapidly over the past few years to provide methods to automatically verify that AI systems robustly behave as intended. This talk will introduce some of the key methods we developed in this area, covering both symbolic and connectionist systems. It will discuss logic-based methods for the verification of unbounded multi-agent systems (i.e., systems composed of an arbitrary number of homogeneous agents, e.g., robot swarms), optimisation approaches for establishing the robustness of neural network models, and methods for analysing properties of neuro-symbolic multi-agent systems.

EC6: Arunesh Sinha

Rutgers Business School

AI and Multi-agent Systems for Real World Decision Making
Game theory is a popular model for studying multi-agent systems. In this talk, I will present my work on modelling of adversarial multi-agent problems using Stackelberg game models, evolving from standard utility maximizing players to rich models of bounded rationality. A first line of work using utility maximizing adversary models introduced the model of audit games and threat screening games, with novel optimization methods for solving these problems at scale. A second line of work has looked at various aspects of learning bounded rational behavior and optimizing strategic decisions based on the learned models; these aspects include study of different classes of bounded rationality, scalable optimization methods for these highly non-linear models, and high fidelity learning of behavior model of multiple interacting agents from data. Overall, data-driven behavior models with principled strategic decision optimization presents many opportunities for research as well as applications for societal benefit.

EC7: Ondrej Kuzelka

Czech Technical University in Prague

Counting and Sampling Models in First-Order Logic
First-order model counting (FOMC) is the task of counting models of a first-order logic sentence over a given set of domain elements. Its weighted variant, WFOMC, generalizes FOMC by assigning weights to the models and has many applications in statistical relational learning. More than ten years of research by various authors has led to identification of non-trivial classes of WFOMC problems that can be solved in time polynomial in the number of domain elements. In this paper, we describe recent works on WFOMC and the related problem of weighted first-order model sampling (WFOMS). We also discuss possible applications of WFOMC and WFOMS within statistical relational learning and beyond, e.g., automated solving of problems from enumerative combinatorics and elementary probability theory. Finally, we mention research problems that still need to be tackled in order to make applications of these methods really practical more broadly.

EC8: Enrico Scala

University of Brescia

AI Planning for Hybrid Systems
When planning the tasks of some physical entities that need to perform actions in the world (e.g., a Robot) it is necessary to take into account quite complex models for ensuring that the found plan is actually executable. This is because the state of the system evolves according to potentially nonlinear dynamics where interdependent discrete and continuous changes happen over the entire course of the task. Systems of this kind are typically compactly represented in planning using languages mixing propositional logic and mathematics. However, these languages are still poorly understood and exploited. What are the actual difficulties for planning in these settings? How can we build systems that can scale up over realistically sized problems? What are the domains which can benefit from these languages? My research aims to answer these questions looking at both theoretical and practical aspects that go from complexity analysis for specific fragments, innovative heuristic search methods and model-to-model compilations. These models and relative planners hold the promise to deliver trustworthy and explainable AI solutions that do not rely on large amounts of data.

EC9: Yang Liu

UC Santa Cruz ByteDance Research

The Importance of Human-Labeled Data in the Era of LLMs
The advent of large language models (LLMs) has brought about a revolution in the development of tailored machine learning models and sparked debates on redefining data requirements. The automation facilitated by the training and implementation of LLMs has led to discussions and aspirations that human-level labeling interventions may no longer hold the same level of importance as in the era of supervised learning.
This paper presents compelling arguments supporting the ongoing relevance of human-labeled data in the era of LLMs.

EC10: Nisarg Shah

University of Toronto

Pushing the Limits of Fairness in Algorithmic Decision-Making
As algorithms and AI models are increasingly used to augment, or even replace, traditional human decision-making, there is a growing interest in ensuring that they treat (groups of) people fairly. While fairness is a relatively new design criterion in many areas of algorithmic decision-making, it has a long history of study in microeconomic theory. In this talk, I will first summarize recent advances that boost fairness guarantees in traditional economic problems such as resource allocation, and then describe how to adapt these fairness notions to novel decision-making paradigms ranging from classification and clustering to recommender systems and conference peer review.

EC11: Aylin Caliskan

University of Washington

Artificial Intelligence, Bias, and Ethics
Although ChatGPT attempts to mitigate bias, when instructed to translate the gender-neutral Turkish sentences “O bir doktor. O bir hemşire” to English, the outcome is biased: “He is a doctor. She is a nurse.” In 2016, we have demonstrated that language representations trained via unsupervised learning automatically embed implicit biases documented in social cognition through the statistical regularities in language corpora. Embedding associations in language, vision, and multi-modal language-vision models reveal that large-scale sociocultural data is a source of implicit human biases regarding gender, race or ethnicity, skin color, ability, age, sexuality, religion, social class, and intersectional associations. The study of gender bias in language, vision, language-vision, and generative AI has highlighted the sexualization of women and girls in AI, while easily accessible generative AI models such as text-to-image generators amplify bias at scale. As AI increasingly automates tasks that determine life’s outcomes and opportunities, the ethics of AI bias has significant implications for human cognition, society, justice, and the future of AI. Thus, it is crucial to advance our understanding of the depth and prevalence of bias in AI to mitigate it both in machines and society.