Learning to Optimize (L2O) stands at the intersection of traditional optimization and machine learning, utilizing the capabilities of machine learning to enhance conventional optimization techniques. As real-world optimization problems frequently share common structures, L2O provides a tool to exploit these structures for better or faster solutions. This tutorial dives deep into L2O techniques, introducing how to accelerate optimization algorithms, promptly estimate the solutions, or even reshape the optimization problem itself, making it more adaptive to real-world applications. By considering the prerequisites for successful applications of L2O and the structure of the optimization problems at hand, this tutorial provides a comprehensive guide for practitioners and researchers alike.

Moreover, the program committee comprised researchers from 12 countries across five continents. The workshop featured six keynote speakers, oral sessions, poster sessions, a panel discussion, and a networking lunch. Of the 20 submitted papers, six were selected for oral and poster presentation, and an additional nine were selected for poster presentation only. The acceptance rate was, therefore, 75%. All accepted papers are published in the open-access workshop's proceedings at https://ceur-ws.org/Vol-3649/.

In medical imaging, particularly in early disease detection and prognosis tasks, discerning the rationale behind an AI model's predictions is crucial for evaluating the reliability of its decisions. Conventional explanation methods face challenges in identifying discernible decisive features in medical image classifications, where discriminative features are subtle or not immediately apparent. To bridge this gap, we propose an explainable model that is equipped with both decision reasoning and feature identification capabilities. Our approach not only detects influential image patterns but also uncovers the decisive features that drive the model's final predictions. By implementing our method, we can efficiently identify and visualise class-specific features leveraged by the data-driven model, providing insights into the decision-making processes of deep learning models. We validated our model in the demanding realm of medical prognosis task, demonstrating its efficacy and potential in enhancing the reliability of AI in healthcare and in discovering new knowledge in diseases where prognostic understanding is limited.

Curriculum learning has been growing in the domain of reinforcement learning as a method of improving training efficiency for various tasks. It involves modifying the difficulty (lessons) of the environment as the agent learns, in order to encourage more optimal agent behavior and higher reward states. However, most curriculum learning methods currently involve discrete transitions of the curriculum or predefined steps by the programmer or using automatic curriculum learning on only a small subset training such as only on an adversary. In this paper, we propose a novel approach to curriculum learning that uses a Setter Model to automatically generate an action space, adversary strength, initialization, and bunching strength. Transportation and traffic optimization is a well known area of study, especially for reinforcement learning based solutions. We specifically look at the bus bunching problem for the context of this study. The main idea of the problem is to minimize the delays caused by inefficient bus timings for passengers arriving and departing from a system of buses. While the heavy exploration in the area makes innovation and improvement with regards to performance marginal, it simultaneously provides an effective baseline for developing new generalized techniques. Our group is particularly interested in examining curriculum learning and its effect on training efficiency and overall performance. We decide to try a lesser known approach to curriculum learning, in which the curriculum is not fixed or discretely thresholded. Our method for automated curriculum learning involves a curriculum that is dynamically chosen and learned by an adversary network made to increase the difficulty of the agent's training, and defined by multiple forms of input. Our results are shown in the following sections of this paper.

The University of Michigan Robotics program focuses on the study of embodied intelligence that must sense, reason, act, and work with people to improve quality of life and productivity equitably across society. ROB 204, part of the core curriculum towards the undergraduate degree in Robotics, introduces students to topics that enable conceptually designing a robotic system to address users' needs from a sociotechnical context. Students are introduced to human-robot interaction (HRI) concepts and the process for socially-engaged design with a Learn-Reinforce-Integrate approach. In this paper, we discuss the course topics and our teaching methodology, and provide recommendations for delivering this material. Overall, students leave the course with a new understanding and appreciation for how human capabilities can inform requirements for a robotics system, how humans can interact with a robot, and how to assess the usability of robotic systems.

Complex planning and scheduling problems have long been solved using various optimization or heuristic approaches. In recent years, imitation learning that aims to learn from expert demonstrations has been proposed as a viable alternative to solving these problems. Generally speaking, imitation learning is designed to learn either the reward (or preference) model or directly the behavioral policy by observing the behavior of an expert. Existing work in imitation learning and inverse reinforcement learning has focused on imitation primarily in unconstrained settings (e.g., no limit on fuel consumed by the vehicle). However, in many real-world domains, the behavior of an expert is governed not only by reward (or preference) but also by constraints. For instance, decisions on self-driving delivery vehicles are dependent not only on the route preferences/rewards (depending on past demand data) but also on the fuel in the vehicle and the time available. In such problems, imitation learning is challenging as decisions are not only dictated by the reward model but are also dependent on a cost-constrained model. In this paper, we provide multiple methods that match expert distributions in the presence of trajectory cost constraints through (a) Lagrangian-based method; (b) Meta-gradients to find a good trade-off between expected return and minimizing constraint violation; and (c) Cost-violation-based alternating gradient. We empirically show that leading imitation learning approaches imitate cost-constrained behaviors poorly and our meta-gradient-based approach achieves the best performance.

Machine-learned language models have transformed everyday life: they steer us when we study, drive, manage money. They have the potential to transform our civilization. But they hallucinate. Their realities are virtual. This note provides a high-level overview of language models and outlines a low-level model of learning machines. It turns out that, after they become capable of recognizing hallucinations and dreaming safely, as humans tend to be, the language-learning machines proceed to generate broader systems of false beliefs and self-confirming theories, as humans tend to do.

We focus on offline imitation learning (IL), which aims to mimic an expert's behavior using demonstrations without any interaction with the environment. One of the main challenges in offline IL is the limited support of expert demonstrations, which typically cover only a small fraction of the state-action space. While it may not be feasible to obtain numerous expert demonstrations, it is often possible to gather a larger set of sub-optimal demonstrations. For example, in treatment optimization problems, there are varying levels of doctor treatments available for different chronic conditions. These range from treatment specialists and experienced general practitioners to less experienced general practitioners. Similarly, when robots are trained to imitate humans in routine tasks, they might learn from individuals with different levels of expertise and efficiency. In this paper, we propose an offline IL approach that leverages the larger set of sub-optimal demonstrations while effectively mimicking expert trajectories. Existing offline IL methods based on behavior cloning or distribution matching often face issues such as overfitting to the limited set of expert demonstrations or inadvertently imitating sub-optimal trajectories from the larger dataset. Our approach, which is based on inverse soft-Q learning, learns from both expert and sub-optimal demonstrations. It assigns higher importance (through learned weights) to aligning with expert demonstrations and lower importance to aligning with sub-optimal ones. A key contribution of our approach, called SPRINQL, is transforming the offline IL problem into a convex optimization over the space of Q functions. Through comprehensive experimental evaluations, we demonstrate that the SPRINQL algorithm achieves state-of-the-art (SOTA) performance on offline IL benchmarks. Code is available at https://github.com/hmhuy2000/SPRINQL.

Direct Preference Optimization (DPO) has recently emerged as a popular approach to improve reinforcement learning with human feedback (RLHF), leading to better techniques to fine-tune large language models (LLM). A weakness of DPO, however, lies in its lack of capability to characterize the diversity of human preferences. Inspired by Mallows' theory of preference ranking, we develop in this paper a new approach, the Mallows-DPO. A distinct feature of this approach is a dispersion index, which reflects the dispersion of human preference to prompts. We show that existing DPO models can be reduced to special cases of this dispersion index, thus unified with Mallows-DPO. More importantly, we demonstrate (empirically) how to use this dispersion index to enhance the performance of DPO in a broad array of benchmark tasks, from synthetic bandit selection to controllable generations and dialogues, while maintaining great generalization capabilities.

Thermal management of power electronics and Electronic Control Units is crucial in times of increasing power densities and limited assembly space. Electric and autonomous vehicles are a prominent application field. Thermal Interface Materials are used to transfer heat from a semiconductor to a heatsink. They are applied along a dispense path onto the semiconductor and spread over its entire surface once the heatsink is joined. To plan this application path, design engineers typically perform an iterative trial-and-error procedure of elaborate simulations and manual experiments. We propose a fully automated optimization approach, which clearly outperforms the current manual path planning and respects all relevant manufacturing constraints. An optimum dispense path increases the reliability of the thermal interface and makes the manufacturing more sustainable by reducing material waste. We show results on multiple real products from automotive series production, including an experimental validation on actual series manufacturing equipment.