Matrix completion is a basic machine learning problem that has wide applications, especially in collaborative filtering and recommender systems. Simple non-convex optimization algorithms are popular and effective in practice. Despite recent progress in proving various non-convex algorithms converge from a good initial point, it remains unclear why random or arbitrary initialization suffices in practice. We prove that the commonly used non-convex objective function for matrix completion has no spurious local minima --- all local minima must also be global. Therefore, many popular optimization algorithms such as (stochastic) gradient descent can provably solve matrix completion with \textit{arbitrary} initialization in polynomial time.

In most online learning algorithms, the weights assigned to the misclassified examples (or support vectors) remain unchanged during the entire learning process. This is clearly insufficient since when a new misclassified example is added to the pool of support vectors, we generally expect it to affect the weights for the existing support vectors. In this paper, we propose a new online learning method, termed Double Updating Online Learning, or DUOL for short. Instead of only assigning a fixed weight to the misclassified example received in current trial, the proposed online learning algorithm also tries to update the weight for one of the existing support vectors. We show that the mistake bound can be significantly improved by the proposed online learning method. Encouraging experimental results show that the proposed technique is in general considerably more effective than the state-of-the-art online learning algorithms.

Knowledge distillation with unlabeled examples is a powerful training paradigm for generating compact and lightweight student models in applications where the amount of labeled data is limited but one has access to a large pool of unlabeled data. In this setting, a large teacher model generates "soft" pseudo-labels for the unlabeled dataset which are then used for training the student model. Despite its success in a wide variety of applications, a shortcoming of this approach is that the teacher's pseudo-labels are often noisy, leading to impaired student performance. In this paper, we present a principled method for knowledge distillation with unlabeled examples that we call Student-Label Mixing (SLaM) and we show that it consistently improves over prior approaches by evaluating it on several standard benchmarks. Finally, we show that SLaM comes with theoretical guarantees; along the way we give an algorithm improving the best-known sample complexity for learning halfspaces with margin under random classification noise, and provide the first convergence analysis for so-called forward loss-adjustment" methods.

In this work, we improve on the upper and lower bounds for the regret of online learning with strongly observable undirected feedback graphs. The best known upper bound for this problem is \mathcal{O}\bigl(\sqrt{\alpha T\ln K}\bigr), where K is the number of actions, \alpha is the independence number of the graph, and T is the time horizon. The \sqrt{\ln K} factor is known to be necessary when \alpha 1 (the experts case). On the other hand, when \alpha K (the bandits case), the minimax rate is known to be \Theta\bigl(\sqrt{KT}\bigr), and a lower bound \Omega\bigl(\sqrt{\alpha T}\bigr) is known to hold for any \alpha . Our improved upper bound \mathcal{O}\bigl(\sqrt{\alpha T(1 \ln(K/\alpha))}\bigr) holds for any \alpha and matches the lower bounds for bandits and experts, while interpolating intermediate cases.

Concretely, sparing more data for validation from training data would limit the performance of the learned model, yet insufficient validation data could result in a sub-optimal selection of the desired model. In this paper, we propose a novel early stopping method called Label Wave, which does not require validation data for selecting the desired model in the presence of label noise. It works by tracking the changes in the model's predictions on the training set during the training process, aiming to halt training before the model unduly fits mislabeled data. This method is empirically supported by our observation that minimum fluctuations in predictions typically occur at the training epoch before the model excessively fits mislabeled data. Through extensive experiments, we show both the effectiveness of the Label Wave method across various settings and its capability to enhance the performance of existing methods for learning with noisy labels. Deep Neural Networks (DNNs) are praised for their remarkable expressive power, which allows them to uncover intricate patterns in high-dimensional data (Montufar et al., 2014; LeCun et al., 2015) and can even fit data with random labels. However, this strength, often termed Memorization (Zhang et al., 2017), can be a double-edged sword, especially when encountering label noise. When label noise exists, the inherent capability of DNNs might cause the model to fit mislabeled examples from noisy datasets, which can deteriorate its generalization performance. Specifically, when DNNs are trained on noisy datasets containing both clean and mislabeled examples, it is often observed that the test error initially decreases and subsequently increases. To prevent DNNs from overconfidently learning from mislabeled examples, many existing methods for learning with noisy labels (Xia et al., 2019; Han et al., 2020; Song et al., 2022; Huang et al., 2023) explicitly or implicitly adopted the operation of halting training before the test error increases--a strategy termed "early stopping". Early stopping relies on model selection, aiming to choose a model that aligns most closely with the true concept from a range of candidate models obtained during the training process (Mohri et al., 2018; Bai et al., 2021). To this end, leveraging hold-out validation data to pinpoint an appropriate early stopping point for model selection becomes a prevalent approach (Xu & Goodacre, 2018) in deep learning. However, this approach heavily relies on additional validation data that is usually derived by splitting the training set, thereby resulting in degraded performance due to insufficient training data.

As artificial intelligence (AI) technologies begin to permeate diverse fields--from healthcare to education--consumers, researchers and policymakers are increasingly raising concerns about whether and how AI is regulated. It is therefore reasonable to anticipate that alignment with principles of'ethical' or'responsible' AI, as well as compliance with law and policy, will form an increasingly important part of AI development. Yet, for the most part, the conventional computer science curriculum is ill-equipped to prepare students for these challenges. To this end, we seek to explore how new educational content related to AI ethics and AI policy can be integrated into both ethics-and technical-focused courses. This paper describes a two-lecture AI policy module that was piloted in a graduate-level introductory machine learning course in 2024. The module, which includes an in-class active learning game, is evaluated using data from student surveys before and after the lectures, and pedagogical motivations and considerations are discussed. We find that the module is successful in engaging otherwise technically-oriented students on the topic of AI policy, increasing student awareness of the social impacts of a variety of AI technologies and developing student interest in the field of AI regulation. Introduction The explosive growth of artificial intelligence (AI) technologies is widely documented and increasingly evident in everyday life: some responses from the search engine Google now include an "AI Overview" inserted before the first webpage link; companies like Tesla and Waymo have seen success in implementing partial or full autonomous driving in vehicles on live roads; and "Apple Intelligence" was the flagship feature for the launch of Apple's new smartphone in fall 2024. Yet what legal or policy response this technological growth will precipitate is less certain [1, 2]. Nevertheless, it should be expected that the development and enactment of regulatory frameworks for AI will demand AI engineers with a command not only of the technical intricacies of AI models, but also of the policy and regulatory landscape for AI development and compliance [3].

Simulation has been pivotal in recent robotics milestones and is poised to play a prominent role in the field's future. However, recent robotic advances often rely on expensive and high-maintenance platforms, limiting access to broader robotics audiences. This work introduces Wheeled Lab, a framework for the low-cost, open-source wheeled platforms that are already widely established in education and research. Through integration with Isaac Lab, Wheeled Lab introduces modern techniques in Sim2Real, such as domain randomization, sensor simulation, and end-to-end learning, to new user communities. To kickstart education and demonstrate the framework's capabilities, we develop three state-of-the-art policies for small-scale RC cars: controlled drifting, elevation traversal, and visual navigation, each trained in simulation and deployed in the real world. By bridging the gap between advanced Sim2Real methods and affordable, available robotics, Wheeled Lab aims to democratize access to cutting-edge tools, fostering innovation and education in a broader robotics context. The full stack, from hardware to software, is low cost and open-source.

Reinforcement learning (RL) agents typically optimize their policies by performing expensive backward passes to update their network parameters. However, some agents can solve new tasks without updating any parameters by simply conditioning on additional context such as their action-observation histories. This paper surveys work on such behavior, known as in-context reinforcement learning.

The capabilities of Large Language Models (LLMs) in low-resource languages lag far behind those in English, making their universal accessibility a significant challenge. To alleviate this, we present $\textit{Franken-Adapter}$, a modular language adaptation approach for decoder-only LLMs with embedding surgery. Our method begins by creating customized vocabularies for target languages and performing language adaptation through embedding tuning on multilingual data. These pre-trained embeddings are subsequently integrated with LLMs that have been instruction-tuned on English alignment data to enable zero-shot cross-lingual transfer. Our experiments on $\texttt{Gemma2}$ models with up to 27B parameters demonstrate improvements of up to 20% across 96 languages, spanning both discriminative and generative tasks, with minimal regressions ($<$1%) in English. Further in-depth analysis reveals the critical role of customizing tokenizers in enhancing language adaptation, while boosting inference efficiency. Additionally, we show the versatility of our method by achieving a 14% improvement over a math-optimized LLM across 20 languages, offering a modular solution to transfer reasoning abilities across languages post hoc.

Prior efforts in building computer-assisted pronunciation training (CAPT) systems often treat automatic pronunciation assessment (APA) and mispronunciation detection and diagnosis (MDD) as separate fronts: the former aims to provide multiple pronunciation aspect scores across diverse linguistic levels, while the latter focuses instead on pinpointing the precise phonetic pronunciation errors made by non-native language learners. However, it is generally expected that a full-fledged CAPT system should perform both functionalities simultaneously and efficiently. In response to this surging demand, we in this work first propose HMamba, a novel CAPT approach that seamlessly integrates APA and MDD tasks in parallel. In addition, we introduce a novel loss function, decoupled cross-entropy loss (deXent), specifically tailored for MDD to facilitate better-supervised learning for detecting mispronounced phones, thereby enhancing overall performance. A comprehensive set of empirical results on the speechocean762 benchmark dataset demonstrates the effectiveness of our approach on APA. Notably, our proposed approach also yields a considerable improvement in MDD performance over a strong baseline, achieving an F1-score of 63.85%. Our codes are made available at https://github.com/Fuann/hmamba