As the number of large language models (LLMs) released to the public grows, there is a pressing need to understand the safety implications associated with these models learning from third-party custom finetuning data. We explore the behavior of LLMs finetuned on noisy custom data containing unsafe content, represented by datasets that contain biases, toxicity, and harmfulness, finding that while aligned LLMs can readily learn this unsafe content, they also tend to forget it more significantly than other examples when subsequently finetuned on safer content. Drawing inspiration from the discrepancies in forgetting, we introduce the "ForgetFilter" algorithm, which filters unsafe data based on how strong the model's forgetting signal is for that data. We demonstrate that the ForgetFilter algorithm ensures safety in customized finetuning without compromising downstream task performance, unlike sequential safety finetuning. ForgetFilter outperforms alternative strategies like replay and moral self-correction in curbing LLMs' ability to assimilate unsafe content during custom finetuning, e.g. 75% lower than not applying any safety measures and 62% lower than using self-correction in toxicity score.

We propose a novel VQA dataset, based on picture stories designed for educating young children, that aims to facilitate comprehensive evaluation and characterization of vision-language models on comprehension tasks. Unlike current VQA datasets that often focus on fact-based memorization and simple reasoning tasks without principled scientific grounding, we collect data containing tasks reflecting different levels of comprehension and underlying cognitive processes, as laid out in Bloom's Taxonomy, a classic framework widely adopted in education research. The proposed BloomVQA dataset can be mapped to a hierarchical graph-based representation of visual stories, enabling automatic data augmentation and novel measures characterizing model consistency across the underlying taxonomy. We demonstrate graded evaluation and reliability analysis based on our proposed consistency metrics on state-of-the-art vision-language models. Our results suggest that, while current models achieve the most gain on low-level comprehension tasks, they generally fall short on high-level tasks requiring more advanced comprehension and cognitive skills, as 38.0% drop in VQA accuracy is observed comparing lowest and highest level tasks. Furthermore, current models show consistency patterns misaligned with human comprehension in various scenarios, suggesting emergent structures of model behaviors.

Federated learning (FL) has shown remarkable success in cooperatively training deep models, while typically struggling with noisy labels. Advanced works propose to tackle label noise by a re-weighting strategy with a strong assumption, i.e., mild label noise. However, it may be violated in many real-world FL scenarios because of highly contaminated clients, resulting in extreme noise ratios, e.g., $>$90%. To tackle extremely noisy clients, we study the robustness of the re-weighting strategy, showing a pessimistic conclusion: minimizing the weight of clients trained over noisy data outperforms re-weighting strategies. To leverage models trained on noisy clients, we propose a novel approach, called negative distillation (FedNed). FedNed first identifies noisy clients and employs rather than discards the noisy clients in a knowledge distillation manner. In particular, clients identified as noisy ones are required to train models using noisy labels and pseudo-labels obtained by global models. The model trained on noisy labels serves as a `bad teacher' in knowledge distillation, aiming to decrease the risk of providing incorrect information. Meanwhile, the model trained on pseudo-labels is involved in model aggregation if not identified as a noisy client. Consequently, through pseudo-labeling, FedNed gradually increases the trustworthiness of models trained on noisy clients, while leveraging all clients for model aggregation through negative distillation. To verify the efficacy of FedNed, we conduct extensive experiments under various settings, demonstrating that FedNed can consistently outperform baselines and achieve state-of-the-art performance. Our code is available at https://github.com/linChen99/FedNed.

Background: We are witnessing an increasing adoption of machine learning (ML), especially deep learning (DL) algorithms in many software systems, including safety-critical systems such as health care systems or autonomous driving vehicles. Ensuring the software quality of these systems is yet an open challenge for the research community, mainly due to the inductive nature of ML software systems. Traditionally, software systems were constructed deductively, by writing down the rules that govern the behavior of the system as program code. However, for ML software, these rules are inferred from training data. Few recent research advances in the quality assurance of ML systems have adapted different concepts from traditional software testing, such as mutation testing, to help improve the reliability of ML software systems. However, it is unclear if any of these proposed testing techniques from research are adopted in practice. There is little empirical evidence about the testing strategies of ML engineers. Aims: To fill this gap, we perform the first fine-grained empirical study on ML testing practices in the wild, to identify the ML properties being tested, the followed testing strategies, and their implementation throughout the ML workflow. Method: First, we systematically summarized the different testing strategies (e.g., Oracle Approximation), the tested ML properties (e.g., Correctness, Bias, and Fairness), and the testing methods (e.g., Unit test) from the literature. Then, we conducted a study to understand the practices of testing ML software. Results: In our findings: 1) we identified four (4) major categories of testing strategy including Grey-box, White-box, Black-box, and Heuristic-based techniques that are used by the ML engineers to find software bugs. 2) We identified 16 ML properties that are tested in the ML workflow.

Motivated by recent work on lifelong learning applications for language models (LMs) of code, we introduce CodeLL, a lifelong learning dataset focused on code changes. Our contribution addresses a notable research gap marked by the absence of a long-term temporal dimension in existing code change datasets, limiting their suitability in lifelong learning scenarios. In contrast, our dataset aims to comprehensively capture code changes across the entire release history of open-source software repositories. In this work, we introduce an initial version of CodeLL, comprising 71 machine-learning-based projects mined from Software Heritage. This dataset enables the extraction and in-depth analysis of code changes spanning 2,483 releases at both the method and API levels. CodeLL enables researchers studying the behaviour of LMs in lifelong fine-tuning settings for learning code changes. Additionally, the dataset can help studying data distribution shifts within software repositories and the evolution of API usages over time.

In the rapidly evolving landscape of education, digital technologies have repeatedly disrupted traditional pedagogical methods. This paper explores the latest of these disruptions: the potential integration of large language models (LLMs) and chatbots into graduate engineering education. We begin by tracing historical and technological disruptions to provide context and then introduce key terms such as machine learning and deep learning and the underlying mechanisms of recent advancements, namely attention/transformer models and graphics processing units. The heart of our investigation lies in the application of an LLM-based chatbot in a graduate fluid mechanics course. We developed a question bank from the course material and assessed the chatbot's ability to provide accurate, insightful responses. The results are encouraging, demonstrating not only the bot's ability to effectively answer complex questions but also the potential advantages of chatbot usage in the classroom, such as the promotion of self-paced learning, the provision of instantaneous feedback, and the reduction of instructors' workload. The study also examines the transformative effect of intelligent prompting on enhancing the chatbot's performance. Furthermore, we demonstrate how powerful plugins like Wolfram Alpha for mathematical problem-solving and code interpretation can significantly extend the chatbot's capabilities, transforming it into a comprehensive educational tool. While acknowledging the challenges and ethical implications surrounding the use of such AI models in education, we advocate for a balanced approach. The use of LLMs and chatbots in graduate education can be greatly beneficial but requires ongoing evaluation and adaptation to ensure ethical and efficient use.

Even though deep speaker models have demonstrated impressive accuracy in speaker verification tasks, this often comes at the expense of increased model size and computation time, presenting challenges for deployment in resource-constrained environments. Our research focuses on addressing this limitation through the development of small footprint deep speaker embedding extraction using knowledge distillation. While previous work in this domain has concentrated on speaker embedding extraction at the utterance level, our approach involves amalgamating embeddings from different levels of the x-vector model (teacher network) to train a compact student network. The results highlight the significance of frame-level information, with the student models exhibiting a remarkable size reduction of 85%-91% compared to their teacher counterparts, depending on the size of the teacher embeddings. Notably, by concatenating teacher embeddings, we achieve student networks that maintain comparable performance to the teacher while enjoying a substantial 75% reduction in model size. These findings and insights extend to other x-vector variants, underscoring the broad applicability of our approach.

As computational systems supported by artificial intelligence (AI) techniques continue to play an increasingly pivotal role in making high-stakes recommendations and decisions across various domains, the demand for explainable AI (XAI) has grown significantly, extending its impact into cognitive learning research. Providing explanations for novel concepts is recognised as a fundamental aid in the learning process, particularly when addressing challenges stemming from knowledge deficiencies and skill application. Addressing these difficulties involves timely explanations and guidance throughout the learning process, prompting the interest of AI experts in developing explainer models. In this paper, we introduce an intelligent system (CL-XAI) for Cognitive Learning which is supported by XAI, focusing on two key research objectives: exploring how human learners comprehend the internal mechanisms of AI models using XAI tools and evaluating the effectiveness of such tools through human feedback. The use of CL-XAI is illustrated with a game-inspired virtual use case where learners tackle combinatorial problems to enhance problem-solving skills and deepen their understanding of complex concepts, highlighting the potential for transformative advances in cognitive learning and co-learning.

Underlying mechanisms of speech perception masked by background speakers, a common daily listening condition, are often investigated using various and lengthy psychophysical tests. The presence of a social agent, such as an interactive humanoid NAO robot, may help maintain engagement and attention. However, such robots potentially have limited sound quality or processing speed. As a first step towards the use of NAO in psychophysical testing of speech-in-speech perception, we compared normal-hearing young adults' performance when using the standard computer interface to that when using a NAO robot to introduce the test and present all corresponding stimuli. Target sentences were presented with colour and number keywords in the presence of competing masker speech at varying target-to-masker ratios. Sentences were produced by the same speaker, but voice differences between the target and masker were introduced using speech synthesis methods. To assess test performance, speech intelligibility and data collection duration were compared between the computer and NAO setups. Human-robot interaction was assessed using the Negative Attitude Towards Robot Scale (NARS) and quantification of behavioural cues (backchannels). Speech intelligibility results showed functional similarity between the computer and NAO setups. Data collection durations were longer when using NAO. NARS results showed participants had a more positive attitude toward robots prior to their interaction with NAO. The presence of more positive backchannels when using NAO suggest higher engagement with the robot in comparison to the computer. Overall, the study presents the potential of the NAO for presentingspeech materials and collecting psychophysical measurements for speech-in-speech perception.

Large language models have shown impressive results for multi-hop mathematical reasoning when the input question is only textual. Many mathematical reasoning problems, however, contain both text and image. With the ever-increasing adoption of vision language models (VLMs), understanding their reasoning abilities for such problems is crucial. In this paper, we evaluate the reasoning capabilities of VLMs along various axes through the lens of geometry problems. We procedurally create a synthetic dataset of geometry questions with controllable difficulty levels along multiple axes, thus enabling a systematic evaluation. The empirical results obtained using our benchmark for state-of-the-art VLMs indicate that these models are not as capable in subjects like geometry (and, by generalization, other topics requiring similar reasoning) as suggested by previous benchmarks. This is made especially clear by the construction of our benchmark at various depth levels, since solving higher-depth problems requires long chains of reasoning rather than additional memorized knowledge. We release the dataset for further research in this area.