As artificial intelligence (AI) technology becomes increasingly prevalent in the filed of education, there is a growing need for mathematics teacher education students (MTES) to demonstrate proficiency in the integration of AI with the technological pedagogical content knowledge (AI-TPACK). To study the issue, we firstly devised an systematic AI-TPACK scale and test on 412 MTES from seven universities. Through descriptive statistical analyses, we found that the current status of AI-TPACK for MTES in China is at a basic, preliminary stage. Secondly, we compared MTES between three different grades on the six variables and found that there is no discernible difference, which suggested that graduate studies were observed to have no promotion in the development of AI-TPACK competencies. Thirdly, we proposed a new AI-TPACK structural equation model (AI-TPACK-SEM) to explore the impact of self-efficacy and teaching beliefs on AI-TPACK. Our findings indicate a positive correlation between self-efficacy and AI-TPACK. We also come to a conclusion that may be contrary to common perception, excessive teaching beliefs may impede the advancement of AI-TPACK. Overall, this paper revealed the current status of AI-TPACK for MTES in China for the first time, designed a dedicated SEM to study the effect of specific factors on AI-TPACK, and proposed some suggestions on future developments.

We study offline reinforcement learning problems with a long-run average reward objective. The state-action pairs generated by any fixed behavioral policy thus follow a Markov chain, and the {\em empirical} state-action-next-state distribution satisfies a large deviations principle. We use the rate function of this large deviations principle to construct an uncertainty set for the unknown {\em true} state-action-next-state distribution. We also construct a distribution shift transformation that maps any distribution in this uncertainty set to a state-action-next-state distribution of the Markov chain generated by a fixed evaluation policy, which may differ from the unknown behavioral policy. We prove that the worst-case average reward of the evaluation policy with respect to all distributions in the shifted uncertainty set provides, in a rigorous statistical sense, the least conservative estimator for the average reward under the unknown true distribution. This guarantee is available even if one has only access to one single trajectory of serially correlated state-action pairs. The emerging robust optimization problem can be viewed as a robust Markov decision process with a non-rectangular uncertainty set. We adapt an efficient policy gradient algorithm to solve this problem. Numerical experiments show that our methods compare favorably against state-of-the-art methods.

Large Language Models (LLMs) have significantly advanced natural language processing with exceptional task generalization capabilities. Low-Rank Adaption (LoRA) offers a cost-effective fine-tuning solution, freezing the original model parameters and training only lightweight, low-rank adapter matrices. However, the memory footprint of LoRA is largely dominated by the original model parameters. To mitigate this, we propose LoRAM, a memory-efficient LoRA training scheme founded on the intuition that many neurons in over-parameterized LLMs have low training utility but are essential for inference. LoRAM presents a unique twist: it trains on a pruned (small) model to obtain pruned low-rank matrices, which are then recovered and utilized with the original (large) model for inference. Additionally, minimal-cost continual pre-training, performed by the model publishers in advance, aligns the knowledge discrepancy between pruned and original models. Our extensive experiments demonstrate the efficacy of LoRAM across various pruning strategies and downstream tasks. For a model with 70 billion parameters, LoRAM enables training on a GPU with only 20G HBM, replacing an A100-80G GPU for LoRA training and 15 GPUs for full fine-tuning. Specifically, QLoRAM implemented by structured pruning combined with 4-bit quantization, for LLaMA-3.1-70B (LLaMA-2-70B), reduces the parameter storage cost that dominates the memory usage in low-rank matrix training by 15.81$\times$ (16.95$\times$), while achieving dominant performance gains over both the original LLaMA-3.1-70B (LLaMA-2-70B) and LoRA-trained LLaMA-3.1-8B (LLaMA-2-13B). Code is available at https://github.com/junzhang-zj/LoRAM.

Offline reinforcement learning (RL) looks at learning how to optimally solve tasks using a fixed dataset of interactions from the environment. Many off-policy algorithms developed for online learning struggle in the offline setting as they tend to over-estimate the behaviour of out of distributions actions. Existing offline RL algorithms adapt off-policy algorithms, employing techniques such as constraining the policy or modifying the value function to achieve good performance on individual datasets but struggle to adapt to different tasks or datasets of different qualities without tuning hyper-parameters. We introduce a policy switching technique that dynamically combines the behaviour of a pure off-policy RL agent, for improving behaviour, and a behavioural cloning (BC) agent, for staying close to the data. We achieve this by using a combination of epistemic uncertainty, quantified by our RL model, and a metric for aleatoric uncertainty extracted from the dataset. We show empirically that our policy switching technique can outperform not only the individual algorithms used in the switching process but also compete with state-of-the-art methods on numerous benchmarks. Our use of epistemic uncertainty for policy switching also allows us to naturally extend our method to the domain of offline to online fine-tuning allowing our model to adapt quickly and safely from online data, either matching or exceeding the performance of current methods that typically require additional modification or hyper-parameter fine-tuning.

Fine-tuning large language models (LLMs) in federated settings enables privacy-preserving adaptation but suffers from cross-client interference due to model aggregation. Existing federated LoRA fine-tuning methods, primarily based on FedAvg, struggle with data heterogeneity, leading to harmful cross-client interference and suboptimal personalization. In this work, we propose \textbf{FedALT}, a novel personalized federated LoRA fine-tuning algorithm that fundamentally departs from FedAvg. Instead of using an aggregated model to initialize local training, each client continues training its individual LoRA while incorporating shared knowledge through a separate Rest-of-the-World (RoTW) LoRA component. To effectively balance local adaptation and global information, FedALT introduces an adaptive mixer that dynamically learns input-specific weightings between the individual and RoTW LoRA components using the Mixture-of-Experts (MoE) principle. Through extensive experiments on NLP benchmarks, we demonstrate that FedALT significantly outperforms state-of-the-art personalized federated LoRA fine-tuning methods, achieving superior local adaptation without sacrificing computational efficiency.

Learning to Optimize (L2O) enhances optimization efficiency with integrated neural networks. L2O paradigms achieve great outcomes, e.g., refitting optimizer, generating unseen solutions iteratively or directly. However, conventional L2O methods require intricate design and rely on specific optimization processes, limiting scalability and generalization. Our analyses explore general framework for learning optimization, called Diff-L2O, focusing on augmenting sampled solutions from a wider view rather than local updates in real optimization process only. Meanwhile, we give the related generalization bound, showing that the sample diversity of Diff-L2O brings better performance. This bound can be simply applied to other fields, discussing diversity, mean-variance, and different tasks. Diff-L2O's strong compatibility is empirically verified with only minute-level training, comparing with other hour-levels.

Accurate fine-grained geospatial scene classification using remote sensing imagery is essential for a wide range of applications. However, existing approaches often rely on manually zooming remote sensing images at different scales to create typical scene samples. This approach fails to adequately support the fixed-resolution image interpretation requirements in real-world scenarios. To address this limitation, we introduce the Million-scale finE-grained geospatial scEne classification dataseT (MEET), which contains over 1.03 million zoom-free remote sensing scene samples, manually annotated into 80 fine-grained categories. In MEET, each scene sample follows a scene-inscene layout, where the central scene serves as the reference, and auxiliary scenes provide crucial spatial context for finegrained classification. Moreover, to tackle the emerging challenge of scene-in-scene classification, we present the Context-Aware Transformer (CAT), a model specifically designed for this task, which adaptively fuses spatial context to accurately classify the scene samples. CAT adaptively fuses spatial context to accurately classify the scene samples by learning attentional features that capture the relationships between the center and auxiliary scenes. Based on MEET, we establish a comprehensive benchmark for fine-grained geospatial scene classification, evaluating CAT against 11 competitive baselines. The results demonstrate that CAT significantly outperforms these baselines, achieving a 1.88% higher balanced accuracy (BA) with the Swin-Large backbone, and a notable 7.87% improvement with the Swin-Huge backbone. Further experiments validate the effectiveness of each module in CAT and show the practical applicability of CAT in the urban functional zone mapping. The source code and dataset will be publicly available at https://jerrywyn.github.io/project/MEET.html.

The rapid development of machine learning (ML) technologies, particularly large language models (LLMs), has led to major advancements in natural language processing (NLP, Abbasi et al. 2023). While much of this advancement happened under the umbrella of the common task framework which espouses transparency and openness (Abbasi et al. 2023), in recent years, closed LLMs such as GPT-3 and GPT-4 have set new performance standards in tasks ranging from text generation to question answering, demonstrating unprecedented capabilities in zero-shot and few-shot learning scenarios (Brown et al. 2020, OpenAI 2023). Given the strong performance of closed LLMs such as GPT-4, many studies within the LLM-as-a-judge paradigm rely on their scores as ground truth benchmarks for evaluating both open and closed LLMs (Chiang and Lee 2023), further entrenching the dominance of SOTA closed LLMs (Vergho et al. 2024). Along with closed LLMs, there are also LLMs where the pre-trained models (i.e., training weights) and inference code are publicly available ("open LLMs") such as Llama (Touvron et al. 2023, Dubey et al. 2024) as well as LLMs where the full training data and training code are also available ("open-source LLMs") such as OLMo (Groeneveld et al. 2024). Open and open-source LLMs provide varying levels of transparency for developers and researchers (Liu et al. 2023). Access to model weights, training data, and inference code enables several benefits for the user-developer-researcher community, including lower costs per input/output token through third-party API services, support for local/offline pre-training and fine-tuning, and deeper analysis of model biases and debiasing strategies. However, the dominance of closed LLMs raises a number of concerns, including accessibility and fairness (Strubell et al. 2020, Bender 2021, Irugalbandara et al. 2024).

Feedback is a very important part the learning process. However, it is challenging to make this feedback both timely and accurate when relying on human markers. This is the challenge that Automated Feedback Generation attempts to address. In this paper, a technique to train such a system on a very small dataset with very long sequences is presented. Both of these attributes make this a very challenging task, however, by using a three stage transfer learning pipeline state-of-the-art results can be achieved with qualitatively accurate but unhuman sounding results. The use of both Automated Essay Scoring and Automated Feedback Generation systems in the real world is also discussed.

Large language models (LLMs) have demonstrated exceptional capabilities in generating text, images, and video content. However, as context length grows, the computational cost of attention increases quadratically with the number of tokens, presenting significant efficiency challenges. This paper presents an analysis of various Key-Value (KV) cache compression strategies, offering a comprehensive taxonomy that categorizes these methods by their underlying principles and implementation techniques. Furthermore, we evaluate their impact on performance and inference latency, providing critical insights into their effectiveness. Our findings highlight the trade-offs involved in KV cache compression and its influence on handling long-context scenarios, paving the way for more efficient LLM implementations.