Automated face mask detection remains an important component of hygiene compli-ance, occupational safety, and public health monitoring, even in post-pandemic envi-ronments where real-time, non-intrusive surveillance is required. Traditional deep learning models offer strong recognition performance but are often impractical for de-ployment on embedded and edge devices due to their computational complexity. Re-cent research has therefore emphasized lightweight and hybrid architectures that maintain high detection accuracy while reducing model size, inference latency, and energy consumption. This review provides an […]

Ver mais

arXiv:2601.14354v1 Announce Type: new Abstract: Joint Embedding Predictive Architectures (JEPA) offer a scalable paradigm for self-supervised learning by predicting latent representations rather than reconstructing high-entropy observations. However, existing formulations rely on textit{deterministic} regression objectives, which mask probabilistic semantics and limit its applicability in stochastic control. In this work, we introduce emph{Variational JEPA (VJEPA)}, a textit{probabilistic} generalization that learns a predictive distribution over future latent states via a variational objective. We show that VJEPA unifies representation learning with Predictive […]

Ver mais

Bitcoin was described as a ‘peer-to-peer electronic cash system’ by its founder, Satoshi Nakamoto, in the Bitcoin whitepaper. However, over the years, the narrative of Bitcoin as ‘Digital Gold’ has become dominant. This is not entirely surprising – for Bitcoin to truly be an alternative monetary system, the store of value function needs to come before the medium of exchange function. But did Bitcoin really live up to its dream of digital gold in 2025? Let’s take a […]

Ver mais

Looking at 30k submissions at a single conference venue and also recent AI written paper with AI written reviews – I’m seriously worried about where this is heading. i decided to pursue a PhD because I really liked working on papers for months, get very interesting clinical findings and then present it really well. But I feel that it is dead now. All recent papers I read in my field are just slops and there is no real […]

Ver mais

arXiv:2505.15175v3 Announce Type: replace Abstract: Backdoor and data-poisoning attacks can flip predictions with tiny training corruptions, yet a sharp theory linking poisoning strength, overparameterization, and regularization is lacking. We analyze ridge least squares with an unpenalized intercept in the high-dimensional regime (p,ntoinfty), (p/nto c). Targeted poisoning is modelled by shifting a (theta)-fraction of one class by a direction (mathbf{v}) and relabelling. Using resolvent techniques and deterministic equivalents from random matrix theory, we derive closed-form limits for the poisoned […]

Ver mais

arXiv:2601.21719v1 Announce Type: cross Abstract: We introduce the (Wishart) projection mechanism, a randomized map of the form $S mapsto M f(S)$ with $M sim W_d(1/r I_d, r)$ and study its differential privacy properties. For vector-valued queries $f$, we prove non-asymptotic DP guarantees without any additive noise, showing that Wishart randomness alone can suffice. For matrix-valued queries, however, we establish a sharp negative result: in the noise-free setting, the mechanism is not DP, and we demonstrate its vulnerability by […]

Ver mais

arXiv:2601.09720v1 Announce Type: new Abstract: Question answering (QA) systems are increasingly deployed across domains. However, their reliability is undermined when retrieved evidence is incomplete, noisy, or uncertain. Existing knowledge graph (KG) based QA frameworks typically represent facts as static and deterministic, failing to capture the evolving nature of information and the uncertainty inherent in reasoning. We present a demonstration of uncertainty-aware dynamic KGs, a framework that combines (i) dynamic construction of evolving KGs, (ii) confidence scoring and uncertainty-aware […]

Ver mais

In this unusually personal essay, Teilhard speaks almost across the table, inviting us to rethink our cosmic standing. He shows how humanity, once dwarfed by vast space and tiny particles, regains significance through a third dimension of reality: complexity. As matter organizes itself into ever-deeper centers, consciousness rises—and in us, becomes self-aware. Humanity is not a cosmic accident, he suggests, but the universe awakening to itself and preparing its next leap.

Ver mais

Ground-level building damage assessment captures critical structural details that remain invisible in satellite imagery, yet this perspective remains severely underexplored in current research. We address this gap by introducing the first war damage dataset for ground-level building segmentation, comprising high-resolution side-view images of war-affected Ukrainian buildings with pixel-wise annotations across six semantic classes: Other, Building, Roof, Damage, Damaged Roof, and Broken Window. To preserve original image resolution while enabling efficient deep learning, we employ a patch-based strategy that […]

Ver mais

This paper introduces a production-grade smart-routing gate-way that boosts permission results by ranking acquirer terminals witha two-stage AI pipeline: a static sieve that enforces business rules andforecasts gateway downtime, followed by a dynamic scorer that computesdecay-weighted, real-time success features and infers per-terminal winprobabilities. The system blends time-window and event-window teleme-try with adaptive half-life revisions and a random-forest classifier, thenorchestrates retries to the next-best route on failure, delivering 4–6%success-rate lift across cards, net-banking, and UPI at Internet scale.Deployed inline […]

Ver mais