CISA Adds Two Known Exploited Vulnerabilities to Catalog
CISA added two vulnerabilities, CVE-2025-68686 (Fortinet) and CVE-2026-16812 (Arista), to its Known Exploited Vulnerabilities Catalog.
CISA Cybersecurity Advisories·27 Jul
Search signals, briefings, benchmarks and glossary terms.
Search signals, briefings, benchmarks and glossary terms.
Use this view to inspect the underlying evidence corpus. For ranked developments, decision posture and interpretation, use Signals.
Raw feed or Signals?
Raw feed is chronological evidence. Signals ranks and interprets material change.
CISA added two vulnerabilities, CVE-2025-68686 (Fortinet) and CVE-2026-16812 (Arista), to its Known Exploited Vulnerabilities Catalog.
MIT Tech Review discusses the shift from chatbots to agentic AI, which executes end-to-end business tasks across workflows and systems.
Truist Financial Corp. appointed Jefferies veteran Craig Mineard as co-head of its Technology Media and Telecom (TMT) investment banking group.
China warns of retaliation if US sanctions Chinese AI firms over alleged improper use of American models for training.
Schneider Electric's VC arm, SE Ventures, is investing in startups across data center, grid resilience, robotics, and industrial AI infrastructure.
OpenAI is expanding its European headquarters in Dublin, creating 250 new jobs, citing growing AI demand.
Bloomberg reports on how companies communicate AI's impact on employment to both investors and employees, balancing efficiency and workforce concerns.
Sir Patrick Vallance will lead a new UK AI taskforce, reporting directly to the Prime Minister, to drive national AI strategy and development.
Nvidia is reportedly orchestrating over $750 billion in new AI deals, sparking concerns about inflated demand and valuations.
HSBC plans to hire over 100 AI specialists and establish a global AI center in Singapore, deepening investment in AI capabilities.
Wired details the individuals shaping AI policy within Donald Trump's inner circle, highlighting a diverse range of views.
CXMT founder Zhu Yiming pledged $5.6 billion in worker bonuses after China IPO, highlighting wealth creation in US-China tech race.
NVIDIA introduced Cosmos-H-Dreams, a framework for real-time generative simulation, specifically for surgical robotics applications.
Chinese chip manufacturer CXMT's market debut saw its stock soar 466%, briefly becoming China's most valuable listed company.
Multiverse Computing SL is raising $570 million at a $1.7 billion valuation to fund efforts in reducing AI operational costs.
OpenAI, Anthropic, Google, and Microsoft significantly increased lobbying expenditure in Washington D.C., signaling growing policy influence efforts.
France questions UK's participation in the EU’s €5bn tech start-up fund amid strained UK-EU relations during 'reset' negotiations.
The Financial Times reports some publishers believe AI will replace authors, creating a premium market for human-written books.
TeamSystem's private equity owners are exploring a stake sale at an €8bn valuation amidst market disruption from AI.
Research proposes detecting talking-face deepfakes by analyzing physiological signals like remote photoplethysmography (rPPG), which are absent in synthetic video.
New research addresses scaling Graph Neural Networks (GNNs) on heterophilic graphs, where prior coarsening-based training methods struggled.
Research identifies optimization collapse and topology as causes preventing Logic Gate Networks (LGNs) from reliably benefiting from increased depth.
Research introduces 'Unbiased Open World Regularization' to address bias in self-supervised learning models and Joint-Embedding Predictive Architectures (JEPAs).
Research proposes patient-aware sampling for pretraining EHR foundation models to address bias from mixed patient data and imbalanced contributions.
CARNet proposes a novel linear-complexity model for multivariate time series forecasting that addresses cross-variate dependencies and periodic patterns.
Research introduces a framework for autoregressive EHR foundation models to incorporate multimodal inputs like ECGs, X-rays, and clinical notes.
Research explores quantum federated learning to enable distributed quantum neural network training without sharing sensitive local data for intelligent services.
Research proposes MA-DAR, a method for continual temporal knowledge graph reasoning that integrates new facts while preserving old knowledge.
IFCLoRA is a new parameter-efficient fine-tuning method for LLMs that optimizes rank allocation across Transformer modules without extra memory or computation.
Research introduces Neural Atom Prevalence (NAP), a Bayesian framework for structured node-level model selection in feedforward neural networks.
Research introduces a parameter-free adaptive sparse attention method using data compression, outperforming fixed patterns and dense attention.
Researchers propose Class-Balanced Softmax, a Bayes Theory-based method, to improve deep learning model performance on imbalanced datasets, addressing limitations of existing methods.
RIS-Kernel introduces a model-agnostic architecture, RIS, reducing LLM self-attention complexity to O(N log N) for long-context inference.
Research analyzes the convergence speed of Low-rank Adaptation (LoRA) for fine-tuning large models, finding exponential oracle calls for $\epsilon$-stationary points.
Research proves ReLU networks with two hidden layers can exactly represent the maximum of up to 10 real numbers using rational linear algebra.
Researchers propose DCS, a unified framework for detecting cross-modal copyright infringement in foundation model outputs by analyzing conditional sensitivity.
Research explores Cross-Domain Off-Policy Evaluation and Learning (OPE/L) for contextual bandits to address few-shot data and new actions in real systems.
Research demonstrates Quasi-Monte Carlo (QMC) initialization improves training convergence in meta-reinforcement learning, outperforming orthogonal defaults.
Research overviews Bayesian and frequentist simulation-based inference with machine learning for inverse problems and parameter estimation.
New research proposes Sharpness-Guided Equilibrium Sampling to address poor generalization in long-tailed learning by combining re-sampling with geometry-aware techniques.
Research explores physically constrained federated additive models (FAMs) for auditable and privacy-preserving SLA-risk prediction in O-RAN networks.
Research identifies assumption gaps in privacy-preserving machine learning auditing protocols, impacting their real-world applicability for accuracy and fairness.
Researchers introduced VRDQ, a variance-reduced distributed Q-learning algorithm for multi-agent reinforcement learning over static and dynamic networks.
Research proposes a method to evaluate the causal impact of ML-assisted decision-making using counterfactual correctness without full RCTs.
Researchers introduced CEL, a new library and benchmark for evaluating counterfactual explanations in explainable AI, focusing on properties like actionability.
Researchers propose SEM-DNN, a heteroscedastic neural simultaneous-equation estimator, to learn bidirectional causal interactions from observational data.
Research explores agentic AI for automated, evidence-grounded root cause analysis of industrial anomalies, addressing explainability and data scarcity.
Hopformer, a two-stage Transformer framework, is introduced for multi-variate time series forecasting by separating common trends from specific information.
HiKV proposes a novel algorithm-hardware co-design to compress the KV cache in LLM decoding, tackling memory bottlenecks for long-context models.
Research proposes General Value Functions (GVF) for remaining useful life (RUL) and failure-mode prediction in predictive maintenance.
© 2026 OneBench: AI Insights. All rights reserved.
Evidence before opinion