Erm
PulseAugur coverage of Erm — every cluster mentioning Erm across labs, papers, and developer communities, ranked by signal.
5 day(s) with sentiment data
-
New research explores optimal aggregation for multiclass classification
This paper explores optimal aggregation techniques for multiclass classification problems, extending existing methods from binary classification. It introduces novel aggregation frameworks that achieve optimal or near-o…
-
AI Training Theory: Focusing on Hardest Examples Accelerates Learning
Researchers have developed a theoretical framework for understanding the effectiveness of data pruning in AI systems, particularly when training on the most challenging examples. The study, published on arXiv, introduce…
-
New research links Galois theory to optimal machine learning strategies
A new research paper explores the theoretical underpinnings of learning group elements from their actions, drawing connections between Galois theory and machine learning concepts like empirical risk minimization (ERM). …
-
New study finds limited benefit for curvature-inspired feature scaling in domain generalization
A recent study evaluated a specific implementation of curvature-inspired scalar feature scaling for domain generalization, named HyperDG. The research found that this method, which rescales Euclidean encoder features du…
-
New method maximizes pAUC using positive-unlabeled data
Researchers have developed a novel method for maximizing the partial area under the receiver operating characteristic curve (pAUC) using only positive and unlabeled data. This approach addresses the common challenge whe…
-
New research advances 'relatively smart learning' with improved sample efficiency
Researchers have advanced the study of "relatively smart learning," a concept where a supervised learner aims to match the performance of any certifiable error guarantee derived from unlabeled data. The new work demonst…
-
New AI calibration method tackles shortcut learning in classifiers
Researchers have proposed a new approach to mitigate shortcut learning in AI classifiers by reframing the problem as one of calibration. Their methods, an in-processing regularizer and a post-hoc recalibration step, aim…
-
Researchers find exponential gap in AI learning complexity between deterministic and randomized methods
Researchers have identified an exponential gap in the complexity of online learning between deterministic and randomized approaches when dealing with thresholds on an unknown order. A study by Attias, Hanneke, and Ramas…
-
New AI model uses sparse routing for retinal pathology analysis
Researchers have developed a new deep learning architecture for analyzing retinal fundus images that utilizes sparse conditional computation. This model pairs a Guided Context Gating (GCG) spatial attention front-end wi…
-
Energy-guided Recursive Model enhances neural network reasoning with Hopfield energies · 2 sources tracked
Researchers have introduced the Energy-guided Recursive Model (ERM), a novel approach to recursive reasoning in neural networks. ERM utilizes explicit Hopfield energies to guide the selection of candidate trajectories, …
-
LLM-PV method enhances program learning efficiency using LLM priors
Researchers have developed a new method called LLM-PV that uses pretrained large language models (LLMs) to improve program learning efficiency. This approach leverages LLMs to generate candidate programs, which are then…
-
New algorithm tackles contaminated and heterogeneous multi-task learning
Researchers have developed a new framework for contaminated multi-task learning, addressing challenges posed by data contamination and heterogeneity across tasks. The proposed method, a filtering-based robust multi-task…
-
New research explores Wasserstein DRO for risk-sensitive estimation and regret optimization
Two new research papers explore the application of Wasserstein distributionally robust optimization (DRO) in different machine learning contexts. The first paper introduces a method for risk-sensitive estimation using W…
-
New combinatorial condition settles proper positive-only learning question
Researchers have settled a long-standing question in machine learning regarding proper positive-only learning. The study establishes that a concept class is properly learnable from positive-only samples if it possesses …
-
Big Pharma Tackles Industrial Heat Decarbonization Amidst Price Volatility
Big Pharma companies like AstraZeneca are increasingly focusing on decarbonizing industrial heat, a significant source of emissions within their supply chains. Geopolitical events and resulting energy price volatility h…