SCPP: A Unified Python Library for Soft Clustering
SCPP is a new open-source Python library that unifies the interface for various soft clustering methods, including fuzzy, probabilistic, and deep learning.
arXiv cs.LG — Machine Learning·23 Jul
Search signals, briefings, company results, benchmarks and glossary terms.
Search signals, briefings, company results, 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.
SCPP is a new open-source Python library that unifies the interface for various soft clustering methods, including fuzzy, probabilistic, and deep learning.
Researchers introduced Directional Kernel Mean Difference (DKMD), a signed statistic for univariate distribution comparison, preserving directional shifts.
Research paper proposes current LLMs suffer from a 'structural monoculture' and future AI systems will be 'systems of systems' for specialized tasks.
Research finds optimizing reward model inference speed in RLHF pipelines significantly reduces bottlenecks, suggesting C++ implementations over PyTorch.
Research explores Gaussian averaging as a smooth surrogate for quantized neural networks, deriving bounds on local oscillation for discontinuous models.
A Good Practice Guide for quantifying uncertainties in machine learning models applied to photoplethysmography (PPG) signals from wearables was published.
New research proposes an algorithm for reinforcement learning that achieves asymptotically optimal regret without dependence on the task horizon.
Research explores expert-guided editing for time-series foundation model forecasts, allowing human feedback to revise model-generated trajectories.
Research explored Time Series Foundation Models (TSFMs) like TimesFM, Chronos, and MOIRAI for zero-shot heart rate variability forecasting from consumer wearables.
New research on the Lattice Deduction Transformer (LDT) reveals neural 'solvers' make one-shot predictions rather than iterative deductions.
Researchers discovered HijackKV, a new security threat in LLM inference that exploits position-independent KV cache reuse for data leakage.
Research explores scaling laws for contrastive representation learning using a sketched linear model, extending existing work on sketched linear regression.
Research introduces Experience Augmented Policy Optimization (EAPO) to improve LLM reasoning via Reinforcement Learning with Verifiable Rewards (RLVR).
Researchers propose Safety-Regulated Adaptive Transfer Reinforcement Learning (SRATRL), a teacher-student framework for safe, efficient domain adaptation.
Researchers propose a neuromorphic-inspired Receptron model for efficient, non-linear classification at the edge, reducing compute and memory.
Research introduces 'in-span learning' for reduced-order models (ROMs), allowing them to self-adapt and maintain accuracy when dynamics drift.
AdvSynGNN proposes a Graph Neural Network architecture designed for robust node-level representation learning in the presence of structural noise and non-homophilous graph topologies.
Research presents a framework to estimate accuracy and communication effort for federated perception models pre-deployment, using intrinsic data properties.
Research explores a geometric perspective on intervention learning, characterizing intervention avoidance in autonomous systems by constraining policies.
Researchers propose a novel differentially private training framework for deep neural networks that selectively privatizes inputs, not labels.
Research explores using language model's internal layer-wise trajectories, beyond maximum softmax probability, to improve uncertainty quantification.
Research provides uniform confidence bands for Kernel Ridge Regression (KRR), improving inferential theory for nonstandard data applications.
Koopman Dreamer proposes a spectrally constrained latent dynamics core for world models to improve stability and control in long imagined trajectories.
STN-TGAT proposes a Transformer-Graph Attention Network for stock ranking and portfolio construction, combining temporal dynamics and cross-sectional dependencies.
Research proposes LAARA, a framework for adaptive rank allocation in parameter-efficient fine-tuning (PEFT), addressing the suboptimality of uniform rank.
New dataset and benchmark, Air Quality Arena, evaluates time-series foundation models for large-scale, multi-region air quality forecasting.
New research proposes Memory Merge DQN, an alternative target network update for Deep Q-networks to improve stability and preserve value function structure.
Research introduces confidence intervals for the ℓ2 Expected Calibration Error (ECE), enhancing statistical evaluation of model calibration.
Research explores if randomness is essential for adaptive data analysis to prevent overfitting and false discovery when reusing datasets.
ArenaRL proposes a tournament-based relative ranking system for training open-ended LLM agents, moving beyond scalar reward models for complex tasks.
Researchers propose SynPre-FL, a framework integrating synthetic data pretraining with federated learning to improve privacy-preserving risk prediction.
New benchmark, CruiseBench, for aero-engine remaining useful life (RUL) prediction uses real-flight profiles, extending N-CMAPSS with more realistic data.
Research presents memory-optimal inference artifacts for transformer attention using Mathematics of Arrays (MoA), improving KV-cache efficiency.
Research explores methods for estimating total variation distance between autoregressive model distributions, accounting for inference optimizations.
Research introduces REGEN, a method for distilling expert LLM knowledge into smaller models using offline reinforcement learning and replay-recycling to reduce compute costs.
Research introduces a framework to audit silent safety failures in tool-augmented LLM agents, focusing on malformed tool responses.
New research challenges current machine unlearning evaluation methods, finding they may favor models that retain forgotten data, proposing a restoration-based audit.
Research on 'Always-On Inter-Cluster Repulsion in Replay-Based Continual Learning' found that geometric separation in feature space may not improve retention.
CoTFormer proposes a recurrent transformer architecture that formalizes Chain-of-Thought as preserving intermediate states for explicit reasoning.
GeoDES, a new diffusion-based model, aims to improve storm-centered weather prediction by augmenting limited regional historical data.
New algorithm for linear quadratic systems achieves anytime regret of order $\mathcal{O}(\sqrt{t})$ with computational efficiency and explicit system dimension dependence.
Research explores the frequentist consistency of Prior-Data Fitted Networks (PFNs) for causal inference, addressing uncertainty quantification.
Research explores leveraging offline supervision to improve reinforcement learning for vision-language-action (VLA) models, addressing out-of-distribution (OOD) performance.
A research paper details large-scale development and validation of a deep learning system, Inspectra CXR, for thoracic disease detection on chest radiographs in Thailand.
Research investigates statistical label fusion for consensus segmentation, finding STAPLE underperforms voting under common conditions.
Research identifies a "Dominant-vs-Dominated" (DvD) imbalance in text-to-image diffusion models where one concept dominates multi-concept generation.
Geometric Attention (GA) proposes a new formalization for transformer attention layers, specifying four independent inputs to define the mechanism.
Research explores information separability in pixel-space diffusion models, finding most information is for small-scale perceptual details.
Research identifies a fundamental 'theory-to-practice gap' in learning with ReLU neural networks and neural operators, related to sampling complexity.
Researchers propose Differentially Private Decoupled Training (DP-DT) to improve utility in differentially private neural network training by decoupling representation learning from privacy enforcement.
© 2026 OneBench: AI Insights. All rights reserved.
Evidence before opinion