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.
Research paper proposes advanced AI/ML models for proactive forecasting of time series network utilization KPIs to optimize resource provisioning.
A Good Practice Guide for quantifying uncertainties in machine learning models applied to photoplethysmography (PPG) signals from wearables was published.
Research explored Time Series Foundation Models (TSFMs) like TimesFM, Chronos, and MOIRAI for zero-shot heart rate variability forecasting from consumer wearables.
Research explores using neural collapse instability to prioritize test cases for Deep Neural Networks, aiming to reduce validation costs in critical domains.
Research evaluates nine pre-trained embedding models for gender bias in ML-based recruitment, analyzing inference from unstructured CV text.
Research introduces DynamicRubric, a method to co-evolve LLM evaluators and policies, addressing score-gap collapse in policy optimization.
Research explores autonomous collaborative learning among Tsetlin Machines (TMs) with consensus-based inference, extending distributed TM learning.
Research presents CURED, a web demonstrator unifying ML-based data cleaning and error models for tabular data.
Research frames Active Inference (AIF) as a convex Markov Decision Process, simplifying policy optimization for expected free energy minimization.
Research explores Gaussian averaging as a smooth surrogate for quantized neural networks, deriving bounds on local oscillation for discontinuous models.
Researchers propose a neuromorphic-inspired Receptron model for efficient, non-linear classification at the edge, reducing compute and memory.
Research details the algorithmic complexity of certified machine unlearning, focusing on optimization accuracy and bounds using convex regularizers.
Research introduces a differentiable penalty, the 'quadrilateral loss,' to quantify and enforce additivity in dense neural networks for interpretability.
Research introduces ELSAA, an efficient low-rank and sparse attention approximation to reduce the quadratic computational cost of Transformers for longer inputs.
Research proposes a hybrid method combining contrastive learning (CoLES) with State Space Models (SSMs) like Mamba for transactional sequences.
Research introduces an interpretable fuzzy rule-based regression extension for the Ex-Fuzzy library to enhance transparency in ML models for regulated domains.
Research explores using classical hardware to accelerate quantum autoencoders for real-time anomaly detection in high-energy physics.
Researchers developed a multi-modal transformer for classifying complex signal patterns from nanopore devices for biomarker identification.
New research proposes Paired Sampling to reduce variance in domain adaptation losses, improving effectiveness in minibatch optimization.
Researchers introduced ARROW, a novel online stochastic variance reduction algorithm for MMD and CORAL loss functions, enabling domain adaptation on streaming data.
FraudShield AI, a hybrid LSTM-Graph Neural Network framework, is proposed for detecting sophisticated financial fraud patterns despite extreme data imbalance.
A research paper benchmarks the performance overhead of confidential computing for GPU-accelerated LLM inference using Intel TDX on NVIDIA H100 hardware.
Research proposes Stochastic Primal-Dual Decoding, a method for generative recommender systems to optimize multiple objectives like fairness and constraints.
Refnd proposes a Relational Generative Process (RGP) for machine learning model evaluation, addressing data leakage in relational datasets.
Research explores Reliability-Aware Hard-Soft Physics-Informed Neural Networks (HSPINN) to robustly solve challenging partial differential equations.
Research identifies that generative foundation models excel at zero-shot forecasting for time series with strong trends, offering an actionable selection rule.
Research explores long-term fairness in AI-driven credit lending, considering dynamic environments, performativity, and socioeconomic outcomes.
Research proposes a method for traceable single-cell data distillation, retaining original cell identifiers to enhance auditability and reuse.
New Bayesian framework integrates input dimension reduction directly into Gaussian process modeling, addressing high-dimensional input challenges.
Researchers propose a time-smoothed proximal linear algorithm and local-regret measure for online optimization of non-convex, non-smooth functions.
© 2026 OneBench: AI Insights. All rights reserved.
Evidence before opinion