Although Large Language Models(LLMs) can generate coherent and contextually relevant text, they often struggle to recognise the intent behind the human user's query. Natural Language Understanding (NLU) models, however, interpret the purpose and key information of user's input to enable responsive interactions. Existing NLU models generally map individual utterances to a dual-level semantic frame, involving sentence-level intent and word-level slot labels. However, real-life conversations primarily consist of multi-turn conversations, involving the interpretation of complex and extended dialogues. Researchers encounter challenges addressing all facets of multi-turn dialogue conversations using a unified single NLU model. This paper introduces a novel approach, MIDAS, leveraging a multi-level intent, domain, and slot knowledge distillation for multi-turn NLU. To achieve this, we construct distinct teachers for varying levels of conversation knowledge, namely, sentence-level intent detection, word-level slot filling, and conversation-level domain classification. These teachers are then fine-tuned to acquire specific knowledge of their designated levels. A multi-teacher loss is proposed to facilitate the combination of these multi-level teachers, guiding a student model in multi-turn dialogue tasks. The experimental results demonstrate the efficacy of our model in improving the overall multi-turn conversation understanding, showcasing the potential for advancements in NLU models through the incorporation of multi-level dialogue knowledge distillation techniques.

We consider the problem of active learning on graphs, which has crucial applications in many real-world networks where labeling node responses is expensive. In this paper, we propose an offline active learning method that selects nodes to query by explicitly incorporating information from both the network structure and node covariates. Building on graph signal recovery theories and the random spectral sparsification technique, the proposed method adopts a two-stage biased sampling strategy that takes both informativeness and representativeness into consideration for node querying. Informativeness refers to the complexity of graph signals that are learnable from the responses of queried nodes, while representativeness refers to the capacity of queried nodes to control generalization errors given noisy node-level information. We establish a theoretical relationship between generalization error and the number of nodes selected by the proposed method. Our theoretical results demonstrate the trade-off between informativeness and representativeness in active learning. Extensive numerical experiments show that the proposed method is competitive with existing graph-based active learning methods, especially when node covariates and responses contain noises. Additionally, the proposed method is applicable to both regression and classification tasks on graphs.

Navigating multiple tasks$\unicode{x2014}$for instance in succession as in continual or lifelong learning, or in distributions as in meta or multi-task learning$\unicode{x2014}$requires some notion of adaptation. Evolution over timescales of millennia has imbued humans and other animals with highly effective adaptive learning and decision-making strategies. Central to these functions are so-called neuromodulatory systems. In this work we introduce an abstract framework for integrating theories and evidence from neuroscience and the cognitive sciences into the design of adaptive artificial reinforcement learning algorithms. We give a concrete instance of this framework built on literature surrounding the neuromodulators Acetylcholine (ACh) and Noradrenaline (NA), and empirically validate the effectiveness of the resulting adaptive algorithm in a non-stationary multi-armed bandit problem. We conclude with a theory-based experiment proposal providing an avenue to link our framework back to efforts in experimental neuroscience.

Future reward estimation is a core component of reinforcement learning agents; i.e., Q-value and state-value functions, predicting an agent's sum of future rewards. Their scalar output, however, obfuscates when or what individual future rewards an agent may expect to receive. We address this by modifying an agent's future reward estimator to predict their next N expected rewards, referred to as Temporal Reward Decomposition (TRD). This unlocks novel explanations of agent behaviour. Through TRD we can: estimate when an agent may expect to receive a reward, the value of the reward and the agent's confidence in receiving it; measure an input feature's temporal importance to the agent's action decisions; and predict the influence of different actions on future rewards. Furthermore, we show that DQN agents trained on Atari environments can be efficiently retrained to incorporate TRD with minimal impact on performance.

Growth in the use of large language models (LLMs) in programming education is altering how students write SQL queries. Traditionally, students relied heavily on web search for coding assistance, but this has shifted with the adoption of LLMs like ChatGPT. However, the comparative process and outcomes of using web search versus LLMs for coding help remain underexplored. To address this, we conducted a randomized interview study in a database classroom to compare web search and LLMs, including a publicly available LLM (ChatGPT) and an instructor-tuned LLM, for writing SQL queries. Our findings indicate that using an instructor-tuned LLM required significantly more interactions than both ChatGPT and web search, but resulted in a similar number of edits to the final SQL query. No significant differences were found in the quality of the final SQL queries between conditions, although the LLM conditions directionally showed higher query quality. Furthermore, students using instructor-tuned LLM reported a lower mental demand. These results have implications for learning and productivity in programming education.

Assessing the capabilities of large language models (LLMs) is often challenging, in part, because it is hard to find tasks to which they have not been exposed during training. We take one step to address this challenge by turning to a new task: focusing on symbolic graphics programs, which are a popular representation for graphics content that procedurally generates visual data. LLMs have shown exciting promise towards program synthesis, but do they understand symbolic graphics programs? Unlike conventional programs, symbolic graphics programs can be translated to graphics content. Here, we characterize an LLM's understanding of symbolic programs in terms of their ability to answer questions related to the graphics content. This task is challenging as the questions are difficult to answer from the symbolic programs alone -- yet, they would be easy to answer from the corresponding graphics content as we verify through a human experiment. To understand symbolic programs, LLMs may need to possess the ability to imagine how the corresponding graphics content would look without directly accessing the rendered visual content. We use this task to evaluate LLMs by creating a large benchmark for the semantic understanding of symbolic graphics programs. This benchmark is built via program-graphics correspondence, hence requiring minimal human efforts. We evaluate current LLMs on our benchmark to elucidate a preliminary assessment of their ability to reason about visual scenes from programs. We find that this task distinguishes existing LLMs and models considered good at reasoning perform better. Lastly, we introduce Symbolic Instruction Tuning (SIT) to improve this ability. Specifically, we query GPT4-o with questions and images generated by symbolic programs. Such data are then used to finetune an LLM. We also find that SIT data can improve the general instruction following ability of LLMs.

Clever Hans, a performing horse, drew amazed crowds wherever he went. With his owner Wilhelm, a maths teacher, he put on incredible displays of arithmetic, beating out the answer to sums with his hooves. Hans even appeared to be able to read, though sceptics insisted the horse was merely responding to signals given by Wilhelm, touring Germany before the First World War. However the trick was done, neither the animal nor the teacher would have been surprised by news this month that horses are more intelligent than previously guessed. Researchers at Nottingham Trent University taught 20 horses to touch cards with their noses in return for treats.

As we assess the risks of AI, we are overlooking a crucial threat. Critics commonly highlight three primary hazards--job disruption, bias, and surveillance/privacy. We hear that AI will cause many people to lose their jobs, from dermatologists to truck drivers to marketers. We hear how AI turns historical correlations into predictions that enforce inequality, so that sentencing algorithms predict more recidivism for Black men than white ones. We hear that apps help authorities watch people, such as Amazon tracking which drivers look away from the road.

The risks posed by Artificial Intelligence (AI) are of considerable concern to academics, auditors, policymakers, AI companies, and the public. However, a lack of shared understanding of AI risks can impede our ability to comprehensively discuss, research, and react to them. This paper addresses this gap by creating an AI Risk Repository to serve as a common frame of reference. This comprises a living database of 777 risks extracted from 43 taxonomies, which can be filtered based on two overarching taxonomies and easily accessed, modified, and updated via our website and online spreadsheets. We construct our Repository with a systematic review of taxonomies and other structured classifications of AI risk followed by an expert consultation. We develop our taxonomies of AI risk using a best-fit framework synthesis. Our high-level Causal Taxonomy of AI Risks classifies each risk by its causal factors (1) Entity: Human, AI; (2) Intentionality: Intentional, Unintentional; and (3) Timing: Pre-deployment; Post-deployment. Our mid-level Domain Taxonomy of AI Risks classifies risks into seven AI risk domains: (1) Discrimination & toxicity, (2) Privacy & security, (3) Misinformation, (4) Malicious actors & misuse, (5) Human-computer interaction, (6) Socioeconomic & environmental, and (7) AI system safety, failures, & limitations. These are further divided into 23 subdomains. The AI Risk Repository is, to our knowledge, the first attempt to rigorously curate, analyze, and extract AI risk frameworks into a publicly accessible, comprehensive, extensible, and categorized risk database. This creates a foundation for a more coordinated, coherent, and complete approach to defining, auditing, and managing the risks posed by AI systems.

Stress is among the most commonly employed quality metrics and optimization criteria for dimension reduction projections of high dimensional data. Complex, high dimensional data is ubiquitous across many scientific disciplines, including machine learning, biology, and the social sciences. One of the primary methods of visualizing these datasets is with two dimensional scatter plots that visually capture some properties of the data. Because visually determining the accuracy of these plots is challenging, researchers often use quality metrics to measure projection accuracy or faithfulness to the full data. One of the most commonly employed metrics, normalized stress, is sensitive to uniform scaling of the projection, despite this act not meaningfully changing anything about the projection. We investigate the effect of scaling on stress and other distance based quality metrics analytically and empirically by showing just how much the values change and how this affects dimension reduction technique evaluations. We introduce a simple technique to make normalized stress scale invariant and show that it accurately captures expected behavior on a small benchmark.