Transformers have significantly impacted domains like natural language processing, computer vision, and robotics, where they improve performance compared to other neural networks. This survey explores how transformers are used in reinforcement learning (RL), where they are seen as a promising solution for addressing challenges such as unstable training, credit assignment, lack of interpretability, and partial observability. We begin by providing a brief domain overview of RL, followed by a discussion on the challenges of classical RL algorithms. Next, we delve into the properties of the transformer and its variants and discuss the characteristics that make them well-suited to address the challenges inherent in RL. We examine the application of transformers to various aspects of RL, including representation learning, transition and reward function modeling, and policy optimization. We also discuss recent research that aims to enhance the interpretability and efficiency of transformers in RL, using visualization techniques and efficient training strategies. Often, the transformer architecture must be tailored to the specific needs of a given application. We present a broad overview of how transformers have been adapted for several applications, including robotics, medicine, language modeling, cloud computing, and combinatorial optimization. We conclude by discussing the limitations of using transformers in RL and assess their potential for catalyzing future breakthroughs in this field.

Automated Machine Learning (AutoML) techniques have recently been introduced to design Collaborative Filtering (CF) models in a data-specific manner. However, existing works either search architectures or hyperparameters while ignoring the fact they are intrinsically related and should be considered together. This motivates us to consider a joint hyperparameter and architecture search method to design CF models. However, this is not easy because of the large search space and high evaluation cost. To solve these challenges, we reduce the space by screening out usefulness yperparameter choices through a comprehensive understanding of individual hyperparameters. Next, we propose a two-stage search algorithm to find proper configurations from the reduced space. In the first stage, we leverage knowledge from subsampled datasets to reduce evaluation costs; in the second stage, we efficiently fine-tune top candidate models on the whole dataset. Extensive experiments on real-world datasets show better performance can be achieved compared with both hand-designed and previous searched models. Besides, ablation and case studies demonstrate the effectiveness of our search framework.

We advance the Cohn-Umans framework for developing fast matrix multiplication algorithms. We introduce, analyze, and search for a new subclass of strong uniquely solvable puzzles (SUSP), which we call simplifiable SUSPs. We show that these puzzles are efficiently verifiable, which remains an open question for general SUSPs. We also show that individual simplifiable SUSPs can achieve the same strength of bounds on the matrix multiplication exponent $\omega$ that infinite families of SUSPs can. We report on the construction, by computer search, of larger SUSPs than previously known for small width. This, combined with our tighter analysis, strengthens the upper bound on the matrix multiplication exponent from $2.66$ to $2.505$ obtainable via this computational approach, and nears the results of the handcrafted constructions of Cohn et al.

Meta-structures are widely used to define which subset of neighbors to aggregate information in heterogeneous information networks (HINs). In this work, we investigate existing meta-structures, including meta-path and meta-graph, and observe that they are initially designed manually with fixed patterns and hence are insufficient to encode various rich semantic information on diverse HINs. Through reflection on their limitation, we define a new concept called meta-multigraph as a more expressive and flexible generalization of meta-graph, and propose a stable differentiable search method to automatically optimize the meta-multigraph for specific HINs and tasks. As the flexibility of meta-multigraphs may propagate redundant messages, we further introduce a complex-to-concise (C2C) meta-multigraph that propagates messages from complex to concise along the depth of meta-multigraph. Moreover, we observe that the differentiable search typically suffers from unstable search and a significant gap between the meta-structures in search and evaluation. To this end, we propose a progressive search algorithm by implicitly narrowing the search space to improve search stability and reduce inconsistency. Extensive experiments are conducted on six medium-scale benchmark datasets and one large-scale benchmark dataset over two representative tasks, i.e., node classification and recommendation. Empirical results demonstrate that our search methods can automatically find expressive meta-multigraphs and C2C meta-multigraphs, enabling our model to outperform state-of-the-art heterogeneous graph neural networks.

In this thesis, we explore the use of complex systems to study learning and adaptation in natural and artificial systems. The goal is to develop autonomous systems that can learn without supervision, develop on their own, and become increasingly complex over time. Complex systems are identified as a suitable framework for understanding these phenomena due to their ability to exhibit growth of complexity. Being able to build learning algorithms that require limited to no supervision would enable greater flexibility and adaptability in various applications. By understanding the fundamental principles of learning in complex systems, we hope to advance our ability to design and implement practical learning algorithms in the future. This thesis makes the following key contributions: the development of a general complexity metric that we apply to search for complex systems that exhibit growth of complexity, the introduction of a coarse-graining method to study computations in large-scale complex systems, and the development of a metric for learning efficiency as well as a benchmark dataset for evaluating the speed of learning algorithms. Our findings add substantially to our understanding of learning and adaptation in natural and artificial systems. Moreover, our approach contributes to a promising new direction for research in this area. We hope these findings will inspire the development of more effective and efficient learning algorithms in the future.

Given a language model (LM), maximum probability is a poor decoding objective for open-ended generation, because it produces short and repetitive text. On the other hand, sampling can often produce incoherent text that drifts from the original topics. We propose contrastive decoding (CD), a reliable decoding approach that optimizes a contrastive objective subject to a plausibility constraint. The contrastive objective returns the difference between the likelihood under a large LM (called the expert, e.g. OPT-13B) and a small LM (called the amateur, e.g. OPT-125M), and the constraint ensures that the outputs are plausible. CD is inspired by the fact that the failures of larger LMs (e.g., repetition, incoherence) are even more prevalent in smaller LMs, and that this difference signals which texts should be preferred. CD requires zero additional training, and produces higher quality text than decoding from the larger LM alone. It also works across model scales (OPT-13B and GPT2-1.5B) and significantly outperforms four strong decoding algorithms (e.g., nucleus, top-k) in automatic and human evaluations across wikipedia, news and story domains.

This paper introduces a local search method for improving an existing program with respect to a measurable objective. Program Optimization with Locally Improving Search (POLIS) exploits the structure of a program, defined by its lines. POLIS improves a single line of the program while keeping the remaining lines fixed, using existing brute-force synthesis algorithms, and continues iterating until it is unable to improve the program's performance. POLIS was evaluated with a 27-person user study, where participants wrote programs attempting to maximize the score of two single-agent games: Lunar Lander and Highway. POLIS was able to substantially improve the participants' programs with respect to the game scores. A proof-of-concept demonstration on existing Stack Overflow code measures applicability in real-world problems. These results suggest that POLIS could be used as a helpful programming assistant for programming problems with measurable objectives.

This paper presents a solution for the problem of optimal planning for a robot in a collaborative human-robot team, where the human supervisor is intermittently available to assist the robot in completing tasks more quickly. Specifically, we address the challenge of computing the fastest path between two configurations in an environment with time constraints on how long the robot can wait for assistance. To solve this problem, we propose a novel approach that utilizes the concepts of budget and critical departure times, which enables us to obtain optimal solutions while scaling to larger problem instances than existing methods. We demonstrate the effectiveness of our approach by comparing it with several baseline algorithms on a city road network and analyzing the quality of the solutions obtained. Our work contributes to the field of robot planning by addressing the critical issue of incorporating human assistance and environmental restrictions, which has significant implications for real-world applications.

Local search algorithms are well-known methods for solving large, hard instances of the satisfiability problem (SAT). The performance of these algorithms crucially depends on heuristics for setting noise parameters and scoring variables. The optimal setting for these heuristics varies for different instance distributions. In this paper, we present an approach for learning effective variable scoring functions and noise parameters by using reinforcement learning. We consider satisfiability problems from different instance distributions and learn specialized heuristics for each of them. Our experimental results show improvements with respect to both a WalkSAT baseline and another local search learned heuristic.

Many robotics applications benefit from being able to compute multiple locally optimal paths in a given configuration space. Examples include path planning for of tethered robots with cable-length constraints, systems involving cables, multi-robot topological exploration & coverage, and, congestion reduction for mobile robots navigation without inter-robot coordination. Existing paradigm is to use topological path planning methods that can provide optimal paths from distinct topological classes available in the underlying configuration space. However, these methods usually require non-trivial and non-universal geometrical constructions, which are prohibitively complex or expensive in 3 or higher dimensional configuration spaces with complex topology. Furthermore, topological methods are unable to distinguish between locally optimal paths that belong to the same topological class but are distinct because of genus-zero obstacles in 3D or due to high-cost or high-curvature regions. In this paper we propose an universal and generalized approach to multiple, locally-optimal path planning using the concept of a novel neighborhood-augmented graph, search-based planning in which can compute paths that are topo-geometrically distinct. This approach can find desired number of locally optimal paths in a wider variety of configuration spaces without requiring any complex pre-processing or geometric constructions. Unlike the existing topological methods, resulting optimal paths are not restricted to distinct topological classes, thus making the algorithm applicable to many other problems where locally optimal and geometrically distinct paths are of interest. We demonstrate the use of our algorithm to planning for shortest traversible paths for a tethered robot in 3D with cable-length constraint, and validate the results in simulations and real robot experimentation.