强化学习（RL）文献的最新进展使机器人主义者能够在模拟环境中自动训练复杂的政策。但是，由于这些方法的样本复杂性差，使用现实世界数据解决强化学习问题仍然是一个具有挑战性的问题。本文介绍了一种新颖的成本整形方法，旨在减少学习稳定控制器所需的样品数量。该方法添加了一个涉及控制Lyapunov功能（CLF）的术语 - 基于模型的控制文献的“能量样”功能 - 到典型的成本配方。理论结果表明，新的成本会导致使用较小的折现因子时稳定控制器，这是众所周知的，以降低样品复杂性。此外，通过确保即使是高度亚最佳的策略也可以稳定系统，添加CLF术语“鲁棒化”搜索稳定控制器。我们通过两个硬件示例演示了我们的方法，在其中我们学习了一个cartpole的稳定控制器和仅使用几秒钟和几分钟的微调数据的A1稳定控制器。

We introduce camouflaged data poisoning attacks, a new attack vector that arises in the context of machine unlearning and other settings when model retraining may be induced. An adversary first adds a few carefully crafted points to the training dataset such that the impact on the model's predictions is minimal. The adversary subsequently triggers a request to remove a subset of the introduced points at which point the attack is unleashed and the model's predictions are negatively affected. In particular, we consider clean-label targeted attacks (in which the goal is to cause the model to misclassify a specific test point) on datasets including CIFAR-10, Imagenette, and Imagewoof. This attack is realized by constructing camouflage datapoints that mask the effect of a poisoned dataset.

Multi-Scale and U-shaped Networks are widely used in various image restoration problems, including deblurring. Keeping in mind the wide range of applications, we present a comparison of these architectures and their effects on image deblurring. We also introduce a new block called as NFResblock. It consists of a Fast Fourier Transformation layer and a series of modified Non-Linear Activation Free Blocks. Based on these architectures and additions, we introduce NFResnet and NFResnet+, which are modified multi-scale and U-Net architectures, respectively. We also use three different loss functions to train these architectures: Charbonnier Loss, Edge Loss, and Frequency Reconstruction Loss. Extensive experiments on the Deep Video Deblurring dataset, along with ablation studies for each component, have been presented in this paper. The proposed architectures achieve a considerable increase in Peak Signal to Noise (PSNR) ratio and Structural Similarity Index (SSIM) value.

Human behavior understanding requires looking at minute details in the large context of a scene containing multiple input modalities. It is necessary as it allows the design of more human-like machines. While transformer approaches have shown great improvements, they face multiple challenges such as lack of data or background noise. To tackle these, we introduce the Forced Attention (FAt) Transformer which utilize forced attention with a modified backbone for input encoding and a use of additional inputs. In addition to improving the performance on different tasks and inputs, the modification requires less time and memory resources. We provide a model for a generalised feature extraction for tasks concerning social signals and behavior analysis. Our focus is on understanding behavior in videos where people are interacting with each other or talking into the camera which simulates the first person point of view in social interaction. FAt Transformers are applied to two downstream tasks: personality recognition and body language recognition. We achieve state-of-the-art results for Udiva v0.5, First Impressions v2 and MPII Group Interaction datasets. We further provide an extensive ablation study of the proposed architecture.

Learned locomotion policies can rapidly adapt to diverse environments similar to those experienced during training but lack a mechanism for fast tuning when they fail in an out-of-distribution test environment. This necessitates a slow and iterative cycle of reward and environment redesign to achieve good performance on a new task. As an alternative, we propose learning a single policy that encodes a structured family of locomotion strategies that solve training tasks in different ways, resulting in Multiplicity of Behavior (MoB). Different strategies generalize differently and can be chosen in real-time for new tasks or environments, bypassing the need for time-consuming retraining. We release a fast, robust open-source MoB locomotion controller, Walk These Ways, that can execute diverse gaits with variable footswing, posture, and speed, unlocking diverse downstream tasks: crouching, hopping, high-speed running, stair traversal, bracing against shoves, rhythmic dance, and more. Video and code release: https://gmargo11.github.io/walk-these-ways/

Large-scale generative models show an impressive ability to perform a wide range of Natural Language Processing (NLP) tasks using in-context learning, where a few examples are used to describe a task to the model. For Machine Translation (MT), these examples are typically randomly sampled from the development dataset with a similar distribution as the evaluation set. However, it is unclear how the choice of these in-context examples and their ordering impacts the output translation quality. In this work, we aim to understand the properties of good in-context examples for MT in both in-domain and out-of-domain settings. We show that the translation quality and the domain of the in-context examples matter and that 1-shot noisy unrelated example can have a catastrophic impact on output quality. While concatenating multiple random examples reduces the effect of noise, a single good prompt optimized to maximize translation quality on the development dataset can elicit learned information from the pre-trained language model. Adding similar examples based on an n-gram overlap with the test source significantly and consistently improves the translation quality of the outputs, outperforming a strong kNN-MT baseline in 2 out of 4 out-of-domain datasets.

State-of-the-art pre-trained language models (PLMs) outperform other models when applied to the majority of language processing tasks. However, PLMs have been found to degrade in performance under distribution shift, a phenomenon that occurs when data at test-time does not come from the same distribution as the source training set. Equally as challenging is the task of obtaining labels in real-time due to issues like long-labeling feedback loops. The lack of adequate methods that address the aforementioned challenges constitutes the need for approaches that continuously adapt the PLM to a distinct distribution. Unsupervised domain adaptation adapts a source model to an unseen as well as unlabeled target domain. While some techniques such as data augmentation can adapt models in several scenarios, they have only been sparsely studied for addressing the distribution shift problem. In this work, we present an approach (MEMO-CL) that improves the performance of PLMs at test-time under distribution shift. Our approach takes advantage of the latest unsupervised techniques in data augmentation and adaptation to minimize the entropy of the PLM's output distribution. MEMO-CL operates on a batch of augmented samples from a single observation in the test set. The technique introduced is unsupervised, domain-agnostic, easy to implement, and requires no additional data. Our experiments result in a 3% improvement over current test-time adaptation baselines.

We are interested in neurosymbolic systems consisting of a high-level symbolic layer for explainable prediction in terms of human-intelligible concepts; and a low-level neural layer for extracting symbols required to generate the symbolic explanation. Real data is often imperfect meaning that even if the symbolic theory remains unchanged, we may still need to address the problem of mapping raw data to high-level symbols, each time there is a change in the data acquisition environment or equipment. Manual (re-)annotation of the raw data each time this happens is laborious and expensive; and automated labelling methods are often imperfect, especially for complex problems. NEUROLOG proposed the use of a semantic loss function that allows an existing feature-based symbolic model to guide the extraction of feature-values from raw data, using `abduction'. However, the experiments demonstrating the use of semantic loss through abduction appear to rely heavily on a domain-specific pre-processing step that enables a prior delineation of feature locations in the raw data. We examine the use of semantic loss in domains where such pre-processing is not possible, or is not obvious. We show that without any prior information about the features, the NEUROLOG approach can continue to predict accurately even with substantially incorrect feature predictions. We show also that prior information about the features in the form of even imperfect pre-training can help correct this situation. These findings are replicated on the original problem considered by NEUROLOG, without the use of feature-delineation. This suggests that symbolic explanations constructed for data in a domain could be re-used in a related domain, by `feature-adaptation' of pre-trained neural extractors using the semantic loss function constrained by abductive feedback.

Recent improvements in conditional generative modeling have made it possible to generate high-quality images from language descriptions alone. We investigate whether these methods can directly address the problem of sequential decision-making. We view decision-making not through the lens of reinforcement learning (RL), but rather through conditional generative modeling. To our surprise, we find that our formulation leads to policies that can outperform existing offline RL approaches across standard benchmarks. By modeling a policy as a return-conditional diffusion model, we illustrate how we may circumvent the need for dynamic programming and subsequently eliminate many of the complexities that come with traditional offline RL. We further demonstrate the advantages of modeling policies as conditional diffusion models by considering two other conditioning variables: constraints and skills. Conditioning on a single constraint or skill during training leads to behaviors at test-time that can satisfy several constraints together or demonstrate a composition of skills. Our results illustrate that conditional generative modeling is a powerful tool for decision-making.

Building an AI agent that can design on its own has been a goal since the 1980s. Recently, deep learning has shown the ability to learn from large-scale data, enabling significant advances in data-driven design. However, learning over prior data limits us only to solve problems that have been solved before and biases data-driven learning towards existing solutions. The ultimate goal for a design agent is the ability to learn generalizable design behavior in a problem space without having seen it before. We introduce a self-learning agent framework in this work that achieves this goal. This framework integrates a deep policy network with a novel tree search algorithm, where the tree search explores the problem space, and the deep policy network leverages self-generated experience to guide the search further. This framework first demonstrates an ability to discover high-performing generative strategies without any prior data, and second, it illustrates a zero-shot generalization of generative strategies across various unseen boundary conditions. This work evaluates the effectiveness and versatility of the framework by solving multiple versions of two engineering design problems without retraining. Overall, this paper presents a methodology to self-learn high-performing and generalizable problem-solving behavior in an arbitrary problem space, circumventing the needs for expert data, existing solutions, and problem-specific learning.