We generalize active learning to address real-world settings with concrete prediction targets where sampling is restricted to an accessible region of the domain, while prediction targets may lie outside this region. We analyze a family of decision rules that sample adaptively to minimize uncertainty about prediction targets. We are the first to show, under general regularity assumptions, that such decision rules converge uniformly to the smallest possible uncertainty obtainable from the accessible data. We demonstrate their strong sample efficiency in two key applications: Active few-shot fine-tuning of large neural networks and safe Bayesian optimization, where they improve significantly upon the state-of-the-art.

Next-generation (xG) wireless networks, with their complex and dynamic nature, present significant challenges to using traditional optimization techniques. Generative AI (GAI) emerges as a powerful tool due to its unique strengths. Unlike traditional optimization techniques and other machine learning methods, GAI excels at learning from real-world network data, capturing its intricacies. This enables safe, offline exploration of various configurations and generation of diverse, unseen scenarios, empowering proactive, data-driven exploration and optimization for xG networks. Additionally, GAI's scalability makes it ideal for large-scale xG networks. This paper surveys how GAI-based models unlock optimization opportunities in xG wireless networks. We begin by providing a review of GAI models and some of the major communication paradigms of xG (e.g., 6G) wireless networks. We then delve into exploring how GAI can be used to improve resource allocation and enhance overall network performance. Additionally, we briefly review the networking requirements for supporting GAI applications in xG wireless networks. The paper further discusses the key challenges and future research directions in leveraging GAI for network optimization. Finally, a case study demonstrates the application of a diffusion-based GAI model for load balancing, carrier aggregation, and backhauling optimization in non-terrestrial networks, a core technology of xG networks. This case study serves as a practical example of how the combination of reinforcement learning and GAI can be implemented to address real-world network optimization problems.

Recent large language models (LLMs) have witnessed significant advancement in various tasks, including mathematical reasoning and theorem proving. As these two tasks require strict and formal multi-step inference, they are appealing domains for exploring the reasoning ability of LLMs but still face important challenges. Previous studies such as Chain-of-Thought (CoT) have revealed the effectiveness of intermediate steps guidance. However, such step-wise annotation requires heavy labor, leading to insufficient training steps for current benchmarks. To fill this gap, this work introduces MUSTARD, a data generation framework that masters uniform synthesis of theorem and proof data of high quality and diversity. MUSTARD synthesizes data in three stages: (1) It samples a few mathematical concept seeds as the problem category. (2) Then, it prompts a generative language model with the sampled concepts to obtain both the problems and their step-wise formal solutions. (3) Lastly, the framework utilizes a proof assistant (e.g., Lean Prover) to filter the valid proofs. With the proposed MUSTARD, we present a theorem-and-proof benchmark MUSTARDSAUCE with 5,866 valid data points. Each data point contains an informal statement, an informal proof, and a translated formal proof that passes the prover validation. We perform extensive analysis and demonstrate that MUSTARD generates validated high-quality step-by-step data. We further apply the MUSTARDSAUCE for fine-tuning smaller language models. The fine-tuned Llama 2-7B achieves a 15.41% average relative performance gain in automated theorem proving, and 8.18% in math word problems. Codes and data are available at https://github.com/Eleanor-H/MUSTARD.

In recent years, with the development of quantum machine learning, quantum neural networks (QNNs) have gained increasing attention in the field of natural language processing (NLP) and have achieved a series of promising results. However, most existing QNN models focus on the architectures of quantum recurrent neural network (QRNN) and self-attention mechanism (QSAM). In this work, we propose a novel QNN model based on quantum convolution. We develop the quantum depthwise convolution that significantly reduces the number of parameters and lowers computational complexity. We also introduce the multi-scale feature fusion mechanism to enhance model performance by integrating word-level and sentence-level features. Additionally, we propose the quantum word embedding and quantum sentence embedding, which provide embedding vectors more efficiently. Through experiments on two benchmark text classification datasets, we demonstrate our model outperforms a wide range of state-of-the-art QNN models. Notably, our model achieves a new state-of-the-art test accuracy of 96.77% on the RP dataset. We also show the advantages of our quantum model over its classical counterparts in its ability to improve test accuracy using fewer parameters. Finally, an ablation test confirms the effectiveness of the multi-scale feature fusion mechanism and quantum depthwise convolution in enhancing model performance.

The recent breakthrough in large language models (LLMs) such as ChatGPT has revolutionized every industry at an unprecedented pace. Alongside this progress also comes mounting concerns about LLMs' susceptibility to jailbreaking attacks, which leads to the generation of harmful or unsafe content. While safety alignment measures have been implemented in LLMs to mitigate existing jailbreak attempts and force them to become increasingly complicated, it is still far from perfect. In this paper, we analyze the common pattern of the current safety alignment and show that it is possible to exploit such patterns for jailbreaking attacks by simultaneous obfuscation in queries and responses. Specifically, we propose WordGame attack, which replaces malicious words with word games to break down the adversarial intent of a query and encourage benign content regarding the games to precede the anticipated harmful content in the response, creating a context that is hardly covered by any corpus used for safety alignment. Extensive experiments demonstrate that WordGame attack can break the guardrails of the current leading proprietary and open-source LLMs, including the latest Claude 3, GPT 4, and Llama 3 models more effectively than existing attacks efficiently. Further ablation studies on such simultaneous obfuscation in query and response provide evidence of the merits of the attack strategy beyond an individual attack. Warning: The paper contains unfiltered text generated by LLMs which can be offensive.

We propose an actor-critic algorithm for a family of complex problems arising in algebraic statistics and discrete optimization. The core task is to produce a sample from a finite subset of the non-negative integer lattice defined by a high-dimensional polytope. We translate the problem into a Markov decision process and devise an actor-critic reinforcement learning (RL) algorithm to learn a set of good moves that can be used for sampling. We prove that the actor-critic algorithm converges to an approximately optimal sampling policy. To tackle complexity issues that typically arise in these sampling problems, and to allow the RL to function at scale, our solution strategy takes three steps: decomposing the starting point of the sample, using RL on each induced subproblem, and reconstructing to obtain a sample in the original polytope. In this setup, the proof of convergence applies to each subproblem in the decomposition. We test the method in two regimes. In statistical applications, a high-dimensional polytope arises as the support set for the reference distribution in a model/data fit test for a broad family of statistical models for categorical data. We demonstrate how RL can be used for model fit testing problems for data sets for which traditional MCMC samplers converge too slowly due to problem size and sparsity structure. To test the robustness of the algorithm and explore its generalization properties, we apply it to synthetically generated data of various sizes and sparsity levels.

Microsoft and non-profit educational organization Khan Academy have formed a partnership that will allow all K-12 educators in the US to access the pilot version of Khanmigo for Teachers at no cost. Khanmigo is an AI-powered teaching assistant that can help teachers find ways to make lessons more fun and engaging. The tool can also quickly create lesson plans and suggest student groups for team activities. Khan Academy says Khanmigo can save teachers an average of five working hours every week. The service previously cost educators 4 a month, but Khan Academy has dropped those fees since its Microsoft partnership allows it to use the Azure OpenAI Service to power Khanmigo for free.

A major challenge facing the world is the provision of equitable and universal access to quality education. Recent advances in generative AI (gen AI) have created excitement about the potential of new technologies to offer a personal tutor for every learner and a teaching assistant for every teacher. The full extent of this dream, however, has not yet materialised. We argue that this is primarily due to the difficulties with verbalising pedagogical intuitions into gen AI prompts and the lack of good evaluation practices, reinforced by the challenges in defining excellent pedagogy. Here we present our work collaborating with learners and educators to translate high level principles from learning science into a pragmatic set of seven diverse educational benchmarks, spanning quantitative, qualitative, automatic and human evaluations; and to develop a new set of fine-tuning datasets to improve the pedagogical capabilities of Gemini, introducing LearnLM-Tutor. Our evaluations show that LearnLM-Tutor is consistently preferred over a prompt tuned Gemini by educators and learners on a number of pedagogical dimensions. We hope that this work can serve as a first step towards developing a comprehensive educational evaluation framework, and that this can enable rapid progress within the AI and EdTech communities towards maximising the positive impact of gen AI in education.

Biased information (recently termed bisinformation) continues to be taught in medical curricula, often long after having been debunked. In this paper, we introduce BRICC, a firstin-class initiative that seeks to mitigate medical bisinformation using machine learning to systematically identify and flag text with potential biases, for subsequent review in an expert-in-the-loop fashion, thus greatly accelerating an otherwise labor-intensive process. A gold-standard BRICC dataset was developed throughout several years, and contains over 12K pages of instructional materials. Medical experts meticulously annotated these documents for bias according to comprehensive coding guidelines, emphasizing gender, sex, age, geography, ethnicity, and race. Using this labeled dataset, we trained, validated, and tested medical bias classifiers. We test three classifier approaches: a binary type-specific classifier, a general bias classifier; an ensemble combining bias type-specific classifiers independently-trained; and a multitask learning (MTL) model tasked with predicting both general and type-specific biases. While MTL led to some improvement on race bias detection in terms of F1-score, it did not outperform binary classifiers trained specifically on each task. On general bias detection, the binary classifier achieves up to 0.923 of AUC, a 27.8% improvement over the baseline. This work lays the foundations for debiasing medical curricula by exploring a novel dataset and evaluating different training model strategies. Hence, it offers new pathways for more nuanced and effective mitigation of bisinformation.

3D spot beamfocusing (SBF), in contrast to conventional angular-domain beamforming, concentrates radiating power within very small volume in both radial and angular domains in the near-field zone. Recently the implementation of channel-state-information (CSI)-independent machine learning (ML)-based approaches have been developed for effective SBF using extremely-largescale-programable-metasurface (ELPMs). These methods involve dividing the ELPMs into subarrays and independently training them with Deep Reinforcement Learning to jointly focus the beam at the Desired Focal Point (DFP). This paper explores near-field SBF using ELPMs, addressing challenges associated with lengthy training times resulting from independent training of subarrays. To achieve a faster CSIindependent solution, inspired by the correlation between the beamfocusing matrices of the subarrays, we leverage transfer learning techniques. First, we introduce a novel similarity criterion based on the Phase Distribution Image of subarray apertures. Then we devise a subarray policy propagation scheme that transfers the knowledge from trained to untrained subarrays. We further enhance learning by introducing Quasi-Liquid-Layers as a revised version of the adaptive policy reuse technique. We show through simulations that the proposed scheme improves the training speed about 5 times. Furthermore, for dynamic DFP management, we devised a DFP policy blending process, which augments the convergence rate up to 8-fold.