tabular data
PulseAugur coverage of tabular data — every cluster mentioning tabular data across labs, papers, and developer communities, ranked by signal.
3 day(s) with sentiment data
-
New method tackles tabular data unlearning challenges
Researchers have introduced Conflict-Aware Unlearning (CAU), a novel method designed to address the unique challenges of machine unlearning in tabular data. Unlike other data types, tabular data presents a 'forget-retai…
-
Foundation models struggle with agricultural data heterogeneity, study finds
A new paper on arXiv explores the challenges of applying foundation models to agriculture, finding that current models struggle with the heterogeneity of agricultural data and landscapes. The research identifies a "pret…
-
New metric scores Transformer attention heads for tabular data
Researchers have developed a new method for scoring the importance of attention heads in Transformer models applied to tabular data. Experiments on 40 datasets showed that removing heads with the lowest importance score…
-
New SeBA framework enhances few-shot learning for tabular data
Researchers have introduced SeBA (Separated-at-Birth Alignment), a novel framework for semi-supervised few-shot learning specifically designed for tabular data. Unlike existing methods that often struggle with defining …
-
New HAPEns method optimizes AI model ensembles for hardware efficiency
Researchers have developed HAPEns, a novel post-hoc ensembling method designed to optimize both predictive performance and hardware efficiency for tabular data. This approach constructs a diverse set of ensembles that l…
-
New deep clustering ensembles tackle imbalanced tabular data
This paper explores the application of unsupervised deep clustering techniques to imbalanced tabular data, a domain where class imbalance typically hinders supervised classification. The researchers introduce two novel …
-
New POSSE-kNN method improves binary classification accuracy
A new machine learning method called POSSE-kNN has been developed for binary classification tasks, particularly for tabular data. This ensemble technique combines bootstrap sampling, random feature subspaces, out-of-bag…
-
XGBoost Outperforms Neural Networks on Tabular Data
Despite the advancements in large language models (LLMs), gradient-boosted tree models like XGBoost continue to outperform neural networks for tabular data. This is attributed to their inherent inductive bias, practical…
-
AI agents learn to debug and fix their own code autonomously
A new research paper introduces CURED, a demonstrator that combines machine learning and database management systems to help users detect, understand, and repair errors in tabular data. Separately, a self-healing AI arc…
-
Foundation models show promise for credit risk prediction in small-data settings
A new research paper explores the application of foundation models, particularly those designed for tabular data, in the field of credit risk prediction. The study benchmarks these newer models against established machi…
-
New IAIML framework enhances interpretable AI for tabular data · 3 sources tracked
Researchers have developed a new framework called Interaction Aware Interpretable Machine Learning (IAIML) designed to improve interpretability in tabular data models. IAIML addresses the limitation of traditional metho…
-
New GAN model tackles class imbalance in tabular data
Researchers have developed ctdGAN, a novel conditional Generative Adversarial Network designed to address class imbalance in tabular datasets. This new model partitions input samples into clusters and employs a probabil…
-
pTNAS accelerates neural architecture search for tabular data
Researchers have developed pTNAS, a novel approach for progressive neural architecture search specifically designed for tabular data. This method efficiently identifies optimal neural network architectures, significantl…
-
New framework offers recourse for LLM tabular data decisions
Researchers have developed a new framework, ASR-ICL, for generating algorithmic recourse in tabular data when using in-context learning (ICL) with large language models. This framework addresses the gap in providing act…
-
New difficulty score enhances tabular data learning reliability
Researchers have developed a new method called Trajectory-based Difficulty Score (TDS) to estimate the difficulty of individual instances in tabular data learning. This score is derived from the cumulative prediction tr…
-
New RTTAD method enhances anomaly detection with risk-aware adaptation
Researchers have developed a new method called RTTAD to improve unsupervised anomaly detection in tabular data, particularly when the definition of 'normal' data shifts over time. The approach uses a dual-task learning …
-
Data Language Models offer native tabular data understanding, outperforming existing methods
Researchers have introduced Data Language Models (DLMs), a new class of foundation models designed to natively understand tabular data without requiring preprocessing. The first DLM, Schema-1, a 140M parameter model tra…