Multimodal-attributed graphs (MAGs) are a fundamental data structure for multimodal graph learning (MGL), enabling both graph-centric and modality-centric tasks. However, our empirical analysis reveals inherent topology quality limitations in real-world MAGs, including noisy interactions, missing connections, and task-agnostic relational structures. A single graph derived from generic relationships is therefore unlikely to be universally optimal for diverse downstream tasks. To address this challenge, we propose Task-aware Modality and Topology co-Evolution (TMTE), a novel MGL framework that jointly and iteratively […]

Ver mais

arXiv:2603.02411v1 Announce Type: new Abstract: Dataset Distillation (DD) compresses large datasets into compact synthetic ones that maintain training performance. However, current methods mainly target sample reduction, with limited consideration of data precision and its impact on efficiency. We propose Quantization-aware Dataset Distillation (QuADD), a unified framework that jointly optimizes dataset compactness and precision under fixed bit budgets. QuADD integrates a differentiable quantization module within the distillation loop, enabling end-to-end co-optimization of synthetic samples and quantization parameters. Guided by […]

Ver mais

arXiv:2601.03329v1 Announce Type: new Abstract: Attention mechanisms represent a fundamental paradigm shift in neural network architectures, enabling models to selectively focus on relevant portions of input sequences through learned weighting functions. This monograph provides a comprehensive and rigorous mathematical treatment of attention mechanisms, encompassing their theoretical foundations, computational properties, and practical implementations in contemporary deep learning systems. Applications in natural language processing, computer vision, and multimodal learning demonstrate the versatility of attention mechanisms. We examine language modeling with […]

Ver mais

Technology is supercharging the attack on democracy and EFF is fighting back. We’re suing to stop government surveillance. We’re fighting to protect free expression online. And we’re building tools to protect your data privacy. Help support our mission with new gear from EFF’s online store, perfect gifts for the digital rights defender in your life. Take 20% your order today with code BLACKFRI. Thanks for being an EFF supporter! Liquid Core Dice are perfect for tabletop games. The metal […]

Ver mais

Spectral optimizers such as Muon have recently shown strong empirical performance in large-scale language model training, but the source and extent of their advantage remain poorly understood. We study this question through the linear associative memory problem, a tractable model for factual recall in transformer-based models. In particular, we go beyond orthogonal embeddings and consider Gaussian inputs and outputs, which allows the number of stored associations to greatly exceed the embedding dimension. Our main result sharply characterizes the […]

Ver mais

For a long time, growth hacking felt like a cheat code. A clever funnel. n A loophole in an algorithm. n A viral trick no one else had noticed yet. If you were early enough, fast enough, or lucky enough, growth followed. But if you’re being honest, you already know the truth: Growth hacking no longer delivers repeatable results. Not because marketers got lazy. Not because audiences disappeared. n But because the environment fundamentally changed. Platforms matured. Algorithms […]

Ver mais

arXiv:2602.12379v1 Announce Type: new Abstract: Estimating longitudinal treatment effects is essential for sequential decision-making but is challenging due to treatment-confounder feedback. While Iterative Conditional Expectation (ICE) G-computation offers a principled approach, its recursive structure suffers from error propagation, corrupting the learned outcome regression models. We propose D3-Net, a framework that mitigates error propagation in ICE training and then applies a robust final correction. First, to interrupt error propagation during learning, we train the ICE sequence using Sequential Doubly […]

Ver mais

arXiv:2603.01951v1 Announce Type: cross Abstract: We study generalized linear prediction under a streaming setting, where each iteration uses only one fresh data point for a gradient-level update. While momentum is well-established in deterministic optimization, a fundamental open question is whether it can accelerate such single-pass non-quadratic stochastic optimization. We propose the first algorithm that successfully incorporates momentum via a novel data-dependent proximal method, achieving dual-momentum acceleration. Our derived excess risk bound decomposes into three components: an improved optimization […]

Ver mais