The transition to serverless architecture is not a buzzword to developers and software developers involved in developer work in macOS. It is a true revolution in the way apps are built, implemented, and scaled. Gone are the days of staring at terminal windows full of endless server logs and sleepless nights in order to keep the infrastructure from hiccuping. Serverless sets that mess with the work of someone. If you have been developing on macOS, then it is […]

Ver mais

The quality of data augmentation serves as a critical determinant for the performance of contrastive learning in EEG tasks. Although this paradigm is promising for utilizing unlabeled data, static or random augmentation strategies often fail to preserve intrinsic information due to the non-stationarity of EEG signals where statistical properties change over time. To address this, we propose RL-BioAug, a framework that leverages a label-efficient reinforcement learning (RL) agent to autonomously determine optimal augmentation policies. While utilizing only a […]

Ver mais

arXiv:2602.20404v1 Announce Type: new Abstract: In tabular Markov decision processes (MDPs) with perfect state observability, each trajectory provides active samples from the transition distributions conditioned on state-action pairs. Consequently, accurate model estimation depends on how the exploration policy allocates visitation frequencies in accordance with the intrinsic complexity of each transition distribution. Building on recent work on coverage-based exploration, we introduce a parameterized family of decomposable and concave objective functions $U_kappa$ that explicitly incorporate both intrinsic estimation complexity and […]

Ver mais

arXiv:2603.02227v1 Announce Type: new Abstract: Can a transformer learn which attention entries matter during training? In principle, yes: attention distributions are highly concentrated, and a small gate network can identify the important entries post-hoc with near-perfect accuracy. In practice, barely. When sparse attention is trained end-to-end, the model’s Q/K/V projections co-adapt to whatever mask is imposed, absorbing the routing signal until learned gates perform little better than frozen random gates. We call this routing absorption and present four […]

Ver mais

arXiv:2602.23516v1 Announce Type: new Abstract: Differentially Private Stochastic Gradient Descent (DP-SGD) is a cornerstone technique for ensuring privacy in deep learning, widely used in both training from scratch and fine-tuning large-scale language models. While DP-SGD predominantly relies on the Gaussian mechanism, the Laplace mechanism remains underutilized due to its reliance on L1 norm clipping. This constraint severely limits its practicality in high-dimensional models because the L1 norm of an n-dimensional gradient can be up to sqrt(n) times larger […]

Ver mais

Prompt-based continual learning methods effectively mitigate catastrophic forgetting. However, most existing methods assign a fixed set of prompts to each task, completely isolating knowledge across tasks and resulting in suboptimal parameter utilization. To address this, we consider the practical needs of continual learning and propose a prompt-sharing framework. This framework constructs a global prompt pool and introduces a task-aware gated routing mechanism that sparsely activates a subset of prompts to achieve dynamic decoupling and collaborative optimization of task-specific […]

Ver mais

arXiv:2603.00702v1 Announce Type: new Abstract: Khmer is a low-resource language characterized by a complex script, presenting significant challenges for optical character recognition (OCR). While document printed text recognition has advanced because of available datasets, performance on other modalities, such as handwritten and scene text, remains limited by data scarcity. Training modality-specific models for each modality does not allow cross-modality transfer learning, from which modalities with limited data could otherwise benefit. Moreover, deploying many modality-specific models results in significant […]

Ver mais

I’ve been running RL agents for portfolio allocation across equities for a while now — daily OHLCV, quarterly fundamentals, TTM metrics, and options surface data as observations. Wanted to share some practical notes on migrating from PPO to SAC since most of the PPO vs SAC discussion I see online is benchmarked on MuJoCo, not financial data. Why PPO stopped being sufficient PPO worked fine on clean single-frequency daily data. The issues showed up when I introduced mixed-frequency […]

Ver mais

This paper clarifies and extends the restricted-variable framework introduced in Drake (2026). Restriction does not alter the underlying analytic object, provided the restricted and unrestricted presentations agree on a nonempty open overlap domain; it changes only the regime of admissible access. In the model case of the completed Riemann zeta function, the paper shows that functional symmetry acts on the argument of the function but does not by itself induce an admissible zero-indexing rule in the restricted Hadamard […]

Ver mais