Grokking refers to a delayed generalization following overfitting when optimizing artificial neural networks with gradient-based methods. In this work, we demonstrate that grokking can be induced by regularization, either explicit or implicit. More precisely, we show that when there exists a model with a property $P$ (e.g., sparse or low-rank weights) that generalizes on the problem of interest, gradient descent with a small but non-zero regularization of $P$ (e.g., $\ell_1$ or nuclear norm regularization) results in grokking. This extends previous work showing that small non-zero weight decay induces grokking. Moreover, our analysis shows that over-parameterization by adding depth makes it possible to grok or ungrok without explicitly using regularization, which is impossible in shallow cases. We further show that the $\ell_2$ norm is not a reliable proxy for generalization when the model is regularized toward a different property $P$, as the $\ell_2$ norm grows in many cases where no weight decay is used, but the model generalizes anyway. We also show that grokking can be amplified solely through data selection, with any other hyperparameter fixed.

Encountering shifted data at test time is a ubiquitous challenge when deploying predictive models. Test-time adaptation (TTA) methods address this issue by continuously adapting a deployed model using only unlabeled test data. While TTA can extend the model's lifespan, it is only a temporary solution. Eventually the model might degrade to the point that it must be taken offline and retrained. To detect such points of ultimate failure, we propose pairing TTA with risk monitoring frameworks that track predictive performance and raise alerts when predefined performance criteria are violated. Specifically, we extend existing monitoring tools based on sequential testing with confidence sequences to accommodate scenarios in which the model is updated at test time and no test labels are available to estimate the performance metrics of interest. Our extensions unlock the application of rigorous statistical risk monitoring to TTA, and we demonstrate the effectiveness of our proposed TTA monitoring framework across a representative set of datasets, distribution shift types, and TTA methods.

Due to the excellent performance in yielding high-quality, zero-shot segmentation, Segment Anything Model (SAM) and its variants have been widely applied in diverse scenarios such as healthcare and intelligent manufacturing. Therefore, effectively compressing SAMs has become an increasingly pressing practical need. In this study, we propose Birkhoff, a novel data-free compression algorithm for SAM and its variants. Unlike quantization, pruning, distillation, and other compression methods, Birkhoff embodies versatility across model types, agility in deployment, faithfulness to the original model, and compactness in model size. Specifically, Birkhoff introduces a novel compression algorithm: Hyper-Compression, whose core principle is to find a dense trajectory to turn a high-dimensional parameter vector into a low-dimensional scalar. Furthermore, Birkhoff designs a dedicated linear layer operator, HyperLinear, to fuse decompression and matrix multiplication to significantly accelerate inference of the compressed SAMs. Extensive experiments on 18 SAMs in the COCO, LVIS, and SA-1B datasets show that Birkhoff performs consistently and competitively in compression time, compression ratio, post-compression performance, and inference speed. For example, Birkhoff can achieve a compression ratio of 5.17x on SAM2-B, with less than 1% performance drop without using any fine-tuning data. Moreover, the compression is finished within 60 seconds for all models.

Background: Public speaking is a vital professional skill, yet it remains a source of significant anxiety for many individuals. Traditional training relies heavily on expert coaching, but recent advances in AI has led to novel types of commercial automated public speaking feedback tools. However, most research has focused on prototypes rather than commercial applications, and little is known about how public speaking experts perceive these tools. Objectives: This study aims to evaluate expert opinions on the efficacy and design of commercial AI-based public speaking training tools and to propose guidelines for their improvement. Methods: The research involved 16 semi-structured interviews and 2 focus groups with public speaking experts. Participants discussed their views on current commercial tools, their potential integration into traditional coaching, and suggestions for enhancing these systems. Results and Conclusions: Experts acknowledged the value of AI tools in handling repetitive, technical aspects of training, allowing coaches to focus on higher-level skills. However they found key issues in current tools, emphasising the need for personalised, understandable, carefully selected feedback and clear instructional design. Overall, they supported a hybrid model combining traditional coaching with AI-supported exercises.

Flooding remains a major global challenge, worsened by climate change and urbanization, demanding advanced solutions for effective disaster management. While traditional 2D flood mapping techniques provide limited insights, 3D flood mapping, powered by deep learning (DL), offers enhanced capabilities by integrating flood extent and depth. This paper presents a comprehensive survey of deep learning-based 3D flood mapping, emphasizing its advancements over 2D maps by integrating flood extent and depth for effective disaster management and urban planning. The survey categorizes deep learning techniques into task decomposition and end-to-end approaches, applicable to both static and dynamic flood features. We compare key DL architectures, highlighting their respective roles in enhancing prediction accuracy and computational efficiency. Additionally, this work explores diverse data sources such as digital elevation models, satellite imagery, rainfall, and simulated data, outlining their roles in 3D flood mapping. The applications reviewed range from real-time flood prediction to long-term urban planning and risk assessment. However, significant challenges persist, including data scarcity, model interpretability, and integration with traditional hydrodynamic models. This survey concludes by suggesting future directions to address these limitations, focusing on enhanced datasets, improved models, and policy implications for flood management. This survey aims to guide researchers and practitioners in leveraging DL techniques for more robust and reliable 3D flood mapping, fostering improved flood management strategies.

We present the design process and findings of the pre-conference workshop at the Machine Learning for Healthcare Conference (2024) entitled Red Teaming Large Language Models for Healthcare, which took place on August 15, 2024. Conference participants, comprising a mix of computational and clinical expertise, attempted to discover vulnerabilities -- realistic clinical prompts for which a large language model (LLM) outputs a response that could cause clinical harm. Red-teaming with clinicians enables the identification of LLM vulnerabilities that may not be recognised by LLM developers lacking clinical expertise. We report the vulnerabilities found, categorise them, and present the results of a replication study assessing the vulnerabilities across all LLMs provided.

Current efforts in AI safety prioritize filtering harmful content, preventing manipulation of human behavior, and eliminating existential risks in cybersecurity or biosecurity. While pressing, this narrow focus overlooks critical human-centric considerations that shape the long-term trajectory of a society. In this position paper, we identify the risks of overlooking the impact of AI on the future of work and recommend comprehensive transition support towards the evolution of meaningful labor with human agency. Through the lens of economic theories, we highlight the intertemporal impacts of AI on human livelihood and the structural changes in labor markets that exacerbate income inequality. Additionally, the closed-source approach of major stakeholders in AI development resembles rent-seeking behavior through exploiting resources, breeding mediocrity in creative labor, and monopolizing innovation. To address this, we argue in favor of a robust international copyright anatomy supported by implementing collective licensing that ensures fair compensation mechanisms for using data to train AI models. We strongly recommend a pro-worker framework of global AI governance to enhance shared prosperity and economic justice while reducing technical debt.

Catastrophic forgetting in deep neural networks occurs when learning new tasks degrades performance on previously learned tasks due to knowledge overwriting. Among the approaches to mitigate this issue, regularization techniques aim to identify and constrain "important" parameters to preserve previous knowledge. In the highly nonconvex optimization landscape of deep learning, we propose a novel perspective: tracking parameters during the final training plateau is more effective than monitoring them throughout the entire training process. We argue that parameters that exhibit higher activity (movement and variability) during this plateau reveal directions in the loss landscape that are relatively flat, making them suitable for adaptation to new tasks while preserving knowledge from previous ones. Our comprehensive experiments demonstrate that this approach achieves superior performance in balancing catastrophic forgetting mitigation with strong performance on newly learned tasks.

We present elsciRL, an open-source Python library to facilitate the application of language solutions on reinforcement learning problems. We demonstrate the potential of our software by extending the Language Adapter with Self-Completing Instruction framework defined in (Osborne, 2024) with the use of LLMs. Our approach can be re-applied to new applications with minimal setup requirements. We provide a novel GUI that allows a user to provide text input for an LLM to generate instructions which it can then self-complete. Empirical results indicate that these instructions \textit{can} improve a reinforcement learning agent's performance. Therefore, we present this work to accelerate the evaluation of language solutions on reward based environments to enable new opportunities for scientific discovery.

--Overfitting in deep neural networks occurs less frequently than expected. This is a puzzling observation, as theory predicts that greater model capacity should eventually lead to overfitting - yet this is rarely seen in practice. But what if overfitting does occur, not globally, but in specific sub-regions of the data space? In this work, we introduce a novel score that measures the forgetting rate of deep models on validation data, capturing what we term local overfitting: a performance degradation confined to certain regions of the input space. We demonstrate that local overfitting can arise even without conventional overfitting, and is closely linked to the double descent phenomenon. Building on these insights, we introduce a two-stage approach that leverages the training history of a single model to recover and retain forgotten knowledge: first, by aggregating checkpoints into an ensemble, and then by distilling it into a single model of the original size, thus enhancing performance without added inference cost. Notably, in the presence of label noise, our method - Knowledge Fusion followed by Knowledge Distillation - outperforms both the original model and independently trained ensembles, achieving a rare win-win scenario: reduced training and inference complexity. Overfitting a training set is considered a fundamental challenge in machine learning. Theoretical analyses predict that as a model gains additional degrees of freedom, its capacity to fit a given training dataset increases. Consequently, there is a point at which the model becomes too specialized for a particular training set, leading to an increase in its generalization error. In deep learning, one would expect to see increased generalization error as the number of parameters and/or training epochs increases. Surprisingly, even vast deep neural networks with billions of parameters seldom adhere to this expectation, and overfitting as a function of epochs is almost never observed [1]. Typically, a significant increase in the number of parameters still results in enhanced performance, or occasionally in peculiar phenomena like the double descent in test error [2], see Section III.