As sequential learning algorithms are increasingly applied to real life, ensuring data privacy while maintaining their utilities emerges as a timely question. In this context, regret minimisation in stochastic bandits under ϵ-global Differential Privacy (DP) has been widely studied. The present literature poses a significant gap between the best-known regret lower and upper bound in this setting, though they "match in order". Thus, we revisit the regret lower and upper bounds of ϵ-global DP bandits and improve both. First, we prove a tighter regret lower bound involving a novel information-theoretic quantity characterising the hardness of ϵ-global DP in stochastic bandits.

Despite its widespread adoption, Adam's advantage over Stochastic Gradient Descent (SGD) lacks a comprehensive theoretical explanation. This paper investigates Adam's sensitivity to rotations of the parameter space. We observe that Adam's performance in training transformers degrades under random rotations of the parameter space, indicating a crucial sensitivity to the choice of basis in practice. This reveals that conventional rotation-invariant assumptions are insufficient to capture Adam's advantages theoretically. To better understand the rotation-dependent properties that benefit Adam, we also identify structured rotations that preserve or even enhance its empirical performance. We then examine the rotation-dependent assumptions in the literature and find that they fall short in explaining Adam's behaviour across various rotation types. In contrast, we verify the orthogonality of the update as a promising indicator of Adam's basis sensitivity, suggesting it may be the key quantity for developing rotation-dependent theoretical frameworks that better explain its empirical success.

Traditional curriculum learning proceeds from easy to hard samples, yet defining a reliable notion of difficulty remains elusive. Prior work has used submodular functions to induce difficulty scores in curriculum learning. We reinterpret adaptive subset selection and formulate it as a multi-armed bandit problem, where each arm corresponds to a submodular function guiding sample selection. We introduce ONLINESUBMOD, a novel online greedy policy that optimizes a utility-driven reward and provably achieves no-regret performance under various sampling regimes. Empirically, ONLINESUBMOD outperforms both traditional curriculum learning and bi-level optimization approaches across vision and language datasets, showing superior accuracy-efficiency tradeoffs. More broadly, we show that validationdriven reward metrics offer a principled way to guide the curriculum schedule. Our code is publicly available at GitHub 2.

Model merging is a technique that combines multiple large pretrained models into a single model, enhancing performance and broadening task adaptability without original data or additional training. However, most existing model merging methods focus primarily on exploring the parameter space, merging models with identical architectures. Despite its potential, merging in the architecture space remains in its early stages due to the vast search space and challenges related to layer compatibility. This paper designs a hierarchical model merging framework named HM3, formulating a bilevel multi-objective model merging problem across both parameter and architecture spaces. At the parameter level, HM3 integrates existing merging methods to quickly identify optimal parameters. Based on these, an actorcritic strategy with efficient policy discretization is employed at the architecture level to explore inference paths with Markov property in the layer-granularity search space for reconstructing these optimal models. By training reusable policy and value networks, HM3 learns Pareto optimal models to provide customized solutions for various tasks. Experimental results on language and vision tasks demonstrate that HM3 outperforms methods focusing solely on the parameter or architecture space.

The long-tail problem in sequential recommender systems stems from imbalanced interaction data, resulting in suboptimal model performance for tail users and items. Recent studies have leveraged head data to enhance tail data for diminish the impact of the long-tail problem. However, these methods often adopt ad-hoc strategies to distinguish between head and tail data, which fails to capture the underlying distributional characteristics and structural properties of each category. Moreover, due to a substantial representational gap exists between head and tail data, head-to-tail enhancement strategies are susceptible to negative transfer, often leading to a decline in overall model performance. To address these issues, we propose a hierarchical partitioning and stepwise enhancement framework, called HPSERec, for long-tailed sequential recommendation. HPSERec partitions the item set into subsets based on a data imbalance metric, assigning an expert network to each subset to capture user-specific local features. Subsequently, we apply knowledge distillation to progressively improve long-tail interest representation, followed by a Sinkhorn optimal transport-based feedback module, which aligns user representations across expert levels through a globally optimal and softly matched mapping. Extensive experiments on three real-world datasets demonstrate that HPSERec consistently outperforms all baseline methods.

Multimodal Dataset Distillation (MDD) seeks to condense large-scale image-text datasets into compact surrogates while retaining their effectiveness for cross-modal learning. Despite recent progress, existing MDD approaches often suffer from Modality Collapse, characterized by over-concentrated intra-modal representations and enlarged distributional gap across modalities. In this paper, for the first time, we identify this issue as stemming from a fundamental conflict between the over-compression behavior inherent in dataset distillation and the cross-modal supervision imposed by contrastive objectives. To alleviate modality collapse, we introduce RepBlend, a novel MDD framework that weakens overdominant cross-modal supervision via representation blending, thereby significantly enhancing intra-modal diversity. Additionally, we observe that current MDD methods impose asymmetric supervision across modalities, resulting in biased optimization. To address this, we propose symmetric projection trajectory matching, which synchronizes the optimization dynamics using modality-specific projection heads, thereby promoting balanced supervision and enhancing cross-modal alignment. Experiments on Flickr-30K and MS-COCO show that RepBlend consistently outperforms prior state-of-the-art MDD methods, achieving significant gains in retrieval performance (e.g., +9.4 IR@10, +6.3 TR@10 under the 100-pair setting) and offering up to 6.7 distillation speedup.

Large language models (LLMs) typically require fine-tuning for domain-specific tasks, and LoRA offers a computationally efficient approach by training lowrank adapters. LoRA is also communication-efficient for federated LLMs when multiple users collaboratively fine-tune a global LLM model without sharing their proprietary raw data.

Language models must be adapted to understand and follow user instructions. Reinforcement learning is widely used to facilitate this - typically using fixed criteria such as "helpfulness" and "harmfulness". In our work, we instead propose using flexible, instruction-specific criteria as a means of broadening the impact that reinforcement learning can have in eliciting instruction following. We propose "Reinforcement Learning from Checklist Feedback" (RLCF). From instructions, we extract checklists and evaluate how well responses satisfy each item--using both AI judges and specialized verifier programs--then combine these scores to compute rewards for RL. We compare RLCF with other alignment methods on top of a strong instruction following model (Qwen2.5-7B-Instruct)

In this paper, we develop a novel algorithm for constructing time-uniform, asymptotic confidence sequences for quantiles under local differential privacy (LDP). The procedure combines dynamically chained parallel stochastic gradient descent (P-SGD) with a randomized response mechanism, thereby guaranteeing privacy protection while simultaneously estimating the target quantile and its variance. A strong Gaussian approximation for the proposed estimator yields asymptotically anytime-valid confidence sequences whose widths obey the law of the iterated logarithm (LIL). Moreover, the method is fully online, offering high computational efficiency and requiring only O(κ)memory, where κdenotes the number of chains and is much smaller than the sample size. Rigorous mathematical proofs and extensive numerical experiments demonstrate the theoretical soundness and practical effectiveness of the algorithm.

Pre-trained stable diffusion models (SD) have shown great advances in visual correspondence. In this paper, we investigate the capabilities of Diffusion Transformers (DiTs) for accurate dense correspondence. Distinct from SD, DiTs exhibit a critical phenomenon in which very few feature activations exhibit significantly larger values than others, known as massive activations, leading to uninformative representations and significant performance degradation for DiTs. The massive activations consistently concentrate at very few fixed dimensions across all image patch tokens, holding little local information. We analyze these dimension-concentrated massive activations and uncover that their concentration is inherently linked to the Adaptive Layer Normalization (AdaLN) in DiTs. Building on these findings, we propose the Diffusion Transformer Feature (DiTF), a training-free AdaLN-based framework that extracts semantically discriminative features from DiTs. Specifically, DiTF leverages AdaLN to adaptively localize and normalize massive activations through channel-wise modulation. Furthermore, a channel discard strategy is introduced to mitigate the adverse effects of massive activations. Experimental results demonstrate that our DiTF outperforms both DINO and SD-based models and establishes a new state-of-the-art performance for DiTs in different visual correspondence tasks (e.g., with +9.4% on Spair-71k and +4.4% on AP-10K-C.S.).