Can MLLMs Detect Phishing? A Comprehensive Security Benchmark Suite Focusing on Dynamic Threats and Multimodal Evaluation in Academic Environmentscs.AI updates on arXiv.org arXiv:2511.15165v1 Announce Type: cross
Abstract: The rapid proliferation of Multimodal Large Language Models (MLLMs) has introduced unprecedented security challenges, particularly in phishing detection within academic environments. Academic institutions and researchers are high-value targets, facing dynamic, multilingual, and context-dependent threats that leverage research backgrounds, academic collaborations, and personal information to craft highly tailored attacks. Existing security benchmarks largely rely on datasets that do not incorporate specific academic background information, making them inadequate for capturing the evolving attack patterns and human-centric vulnerability factors specific to academia. To address this gap, we present AdapT-Bench, a unified methodological framework and benchmark suite for systematically evaluating MLLM defense capabilities against dynamic phishing attacks in academic settings.
arXiv:2511.15165v1 Announce Type: cross
Abstract: The rapid proliferation of Multimodal Large Language Models (MLLMs) has introduced unprecedented security challenges, particularly in phishing detection within academic environments. Academic institutions and researchers are high-value targets, facing dynamic, multilingual, and context-dependent threats that leverage research backgrounds, academic collaborations, and personal information to craft highly tailored attacks. Existing security benchmarks largely rely on datasets that do not incorporate specific academic background information, making them inadequate for capturing the evolving attack patterns and human-centric vulnerability factors specific to academia. To address this gap, we present AdapT-Bench, a unified methodological framework and benchmark suite for systematically evaluating MLLM defense capabilities against dynamic phishing attacks in academic settings. Read More
The U.S. Securities and Exchange Commission (SEC) has abandoned its lawsuit against SolarWinds and its chief information security officer, alleging that the company had misled investors about the security practices that led to the 2020 supply chain attack. In a joint motion filed November 20, 2025, the SEC, along with SolarWinds and its CISO Timothy […]
Enabling MoE on the Edge via Importance-Driven Expert Schedulingcs.AI updates on arXiv.org arXiv:2508.18983v2 Announce Type: replace
Abstract: The Mixture of Experts (MoE) architecture has emerged as a key technique for scaling Large Language Models by activating only a subset of experts per query. Deploying MoE on consumer-grade edge hardware, however, is constrained by limited device memory, making dynamic expert offloading essential. Unlike prior work that treats offloading purely as a scheduling problem, we leverage expert importance to guide decisions, substituting low-importance activated experts with functionally similar ones already cached in GPU memory, thereby preserving accuracy. As a result, this design reduces memory usage and data transfer, while largely eliminating PCIe overhead. In addition, we introduce a scheduling policy that maximizes the reuse ratio of GPU-cached experts, further boosting efficiency. Extensive evaluations show that our approach delivers 48% lower decoding latency with over 60% expert cache hit rate, while maintaining nearly lossless accuracy.
arXiv:2508.18983v2 Announce Type: replace
Abstract: The Mixture of Experts (MoE) architecture has emerged as a key technique for scaling Large Language Models by activating only a subset of experts per query. Deploying MoE on consumer-grade edge hardware, however, is constrained by limited device memory, making dynamic expert offloading essential. Unlike prior work that treats offloading purely as a scheduling problem, we leverage expert importance to guide decisions, substituting low-importance activated experts with functionally similar ones already cached in GPU memory, thereby preserving accuracy. As a result, this design reduces memory usage and data transfer, while largely eliminating PCIe overhead. In addition, we introduce a scheduling policy that maximizes the reuse ratio of GPU-cached experts, further boosting efficiency. Extensive evaluations show that our approach delivers 48% lower decoding latency with over 60% expert cache hit rate, while maintaining nearly lossless accuracy. Read More
An Implementation of Fully Traced and Evaluated Local LLM Pipeline Using Opik for Transparent, Measurable, and Reproducible AI WorkflowsMarkTechPost In this tutorial, we implement a complete workflow for building, tracing, and evaluating an LLM pipeline using Opik. We structure the system step-by-step, beginning with a lightweight model, adding prompt-based planning, creating a dataset, and finally running automated evaluations. As we move through each snippet, we see how Opik helps us track every function span,
The post An Implementation of Fully Traced and Evaluated Local LLM Pipeline Using Opik for Transparent, Measurable, and Reproducible AI Workflows appeared first on MarkTechPost.
In this tutorial, we implement a complete workflow for building, tracing, and evaluating an LLM pipeline using Opik. We structure the system step-by-step, beginning with a lightweight model, adding prompt-based planning, creating a dataset, and finally running automated evaluations. As we move through each snippet, we see how Opik helps us track every function span,
The post An Implementation of Fully Traced and Evaluated Local LLM Pipeline Using Opik for Transparent, Measurable, and Reproducible AI Workflows appeared first on MarkTechPost. Read More
ChatGPT group chats may help teams bring AI into daily planningAI News OpenAI has introduced group chats inside ChatGPT, giving people a way to bring up to 20 others into a shared conversation with the chatbot. The feature is now available to all logged-in users after a short pilot earlier this month, and it shifts ChatGPT from a mostly one-on-one tool to something that supports small-group collaboration.
The post ChatGPT group chats may help teams bring AI into daily planning appeared first on AI News.
OpenAI has introduced group chats inside ChatGPT, giving people a way to bring up to 20 others into a shared conversation with the chatbot. The feature is now available to all logged-in users after a short pilot earlier this month, and it shifts ChatGPT from a mostly one-on-one tool to something that supports small-group collaboration.
The post ChatGPT group chats may help teams bring AI into daily planning appeared first on AI News. Read More
MMG: Mutual Information Estimation via the MMSE Gap in Diffusioncs.AI updates on arXiv.org arXiv:2509.20609v2 Announce Type: replace-cross
Abstract: Mutual information (MI) is one of the most general ways to measure relationships between random variables, but estimating this quantity for complex systems is challenging. Denoising diffusion models have recently set a new bar for density estimation, so it is natural to consider whether these methods could also be used to improve MI estimation. Using the recently introduced information-theoretic formulation of denoising diffusion models, we show the diffusion models can be used in a straightforward way to estimate MI. In particular, the MI corresponds to half the gap in the Minimum Mean Square Error (MMSE) between conditional and unconditional diffusion, integrated over all Signal-to-Noise-Ratios (SNRs) in the noising process. Our approach not only passes self-consistency tests but also outperforms traditional and score-based diffusion MI estimators. Furthermore, our method leverages adaptive importance sampling to achieve scalable MI estimation, while maintaining strong performance even when the MI is high.
arXiv:2509.20609v2 Announce Type: replace-cross
Abstract: Mutual information (MI) is one of the most general ways to measure relationships between random variables, but estimating this quantity for complex systems is challenging. Denoising diffusion models have recently set a new bar for density estimation, so it is natural to consider whether these methods could also be used to improve MI estimation. Using the recently introduced information-theoretic formulation of denoising diffusion models, we show the diffusion models can be used in a straightforward way to estimate MI. In particular, the MI corresponds to half the gap in the Minimum Mean Square Error (MMSE) between conditional and unconditional diffusion, integrated over all Signal-to-Noise-Ratios (SNRs) in the noising process. Our approach not only passes self-consistency tests but also outperforms traditional and score-based diffusion MI estimators. Furthermore, our method leverages adaptive importance sampling to achieve scalable MI estimation, while maintaining strong performance even when the MI is high. Read More
Mathematical Analysis of Hallucination Dynamics in Large Language Models: Uncertainty Quantification, Advanced Decoding, and Principled Mitigationcs.AI updates on arXiv.org arXiv:2511.15005v1 Announce Type: cross
Abstract: Large Language Models (LLMs) are powerful linguistic engines but remain susceptible to hallucinations: plausible-sounding outputs that are factually incorrect or unsupported. In this work, we present a mathematically grounded framework to understand, measure, and mitigate these hallucinations. Drawing on probabilistic modeling, information theory, trigonometric signal analysis, and Bayesian uncertainty estimation, we analyze how errors compound autoregressively, propose refined uncertainty metrics, including semantic and phase-aware variants, and develop principled mitigation strategies such as contrastive decoding, retrieval-augmented grounding, factual alignment, and abstention. This unified lens connects recent advances in calibration, retrieval, and alignment to support safer and more reliable LLMs.
arXiv:2511.15005v1 Announce Type: cross
Abstract: Large Language Models (LLMs) are powerful linguistic engines but remain susceptible to hallucinations: plausible-sounding outputs that are factually incorrect or unsupported. In this work, we present a mathematically grounded framework to understand, measure, and mitigate these hallucinations. Drawing on probabilistic modeling, information theory, trigonometric signal analysis, and Bayesian uncertainty estimation, we analyze how errors compound autoregressively, propose refined uncertainty metrics, including semantic and phase-aware variants, and develop principled mitigation strategies such as contrastive decoding, retrieval-augmented grounding, factual alignment, and abstention. This unified lens connects recent advances in calibration, retrieval, and alignment to support safer and more reliable LLMs. Read More
STREAM-VAE: Dual-Path Routing for Slow and Fast Dynamics in Vehicle Telemetry Anomaly Detectioncs.AI updates on arXiv.org arXiv:2511.15339v1 Announce Type: cross
Abstract: Automotive telemetry data exhibits slow drifts and fast spikes, often within the same sequence, making reliable anomaly detection challenging. Standard reconstruction-based methods, including sequence variational autoencoders (VAEs), use a single latent process and therefore mix heterogeneous time scales, which can smooth out spikes or inflate variances and weaken anomaly separation.
In this paper, we present STREAM-VAE, a variational autoencoder for anomaly detection in automotive telemetry time-series data. Our model uses a dual-path encoder to separate slow drift and fast spike signal dynamics, and a decoder that represents transient deviations separately from the normal operating pattern. STREAM-VAE is designed for deployment, producing stable anomaly scores across operating modes for both in-vehicle monitors and backend fleet analytics.
Experiments on an automotive telemetry dataset and the public SMD benchmark show that explicitly separating drift and spike dynamics improves robustness compared to strong forecasting, attention, graph, and VAE baselines.
arXiv:2511.15339v1 Announce Type: cross
Abstract: Automotive telemetry data exhibits slow drifts and fast spikes, often within the same sequence, making reliable anomaly detection challenging. Standard reconstruction-based methods, including sequence variational autoencoders (VAEs), use a single latent process and therefore mix heterogeneous time scales, which can smooth out spikes or inflate variances and weaken anomaly separation.
In this paper, we present STREAM-VAE, a variational autoencoder for anomaly detection in automotive telemetry time-series data. Our model uses a dual-path encoder to separate slow drift and fast spike signal dynamics, and a decoder that represents transient deviations separately from the normal operating pattern. STREAM-VAE is designed for deployment, producing stable anomaly scores across operating modes for both in-vehicle monitors and backend fleet analytics.
Experiments on an automotive telemetry dataset and the public SMD benchmark show that explicitly separating drift and spike dynamics improves robustness compared to strong forecasting, attention, graph, and VAE baselines. Read More
China-linked APT24 hackers have been using a previously undocumented malware called BadAudio in a three-year espionage campaign that recently switched to more sophisticated attack methods. […] Read More
Cybersecurity researchers have warned of an actively expanding botnet dubbed Tsundere that’s targeting Windows users. Active since mid-2025, the threat is designed to execute arbitrary JavaScript code retrieved from a command-and-control (C2) server, Kaspersky researcher Lisandro Ubiedo said in an analysis published today. There are currently no details on how the botnet malware is propagated; Read […]
