AutoLINC automates loss function design for better handling of class imbalance in machine learning.
― 7 min read
Computer Science - Artificial Intelligence
RSSA new method for assessing language models' alignment with human values.
― 7 min read
A fresh look at how facts are stored in language models.
― 5 min read
Addressing data augmentation issues for better performance in Vision Transformers.
― 5 min read
Discover how Exclusively Penalized Q-learning improves offline RL learning and performance.
― 6 min read
A new approach improves the security of neural networks against adversarial examples.
― 6 min read
Research shows simple classifiers can spot unique patterns in AI-generated text.
― 6 min read
A new method enhances out-of-distribution detection for AI in math tasks.
― 5 min read
ClimODE combines machine learning and physics for improved weather predictions.
― 4 min read
Exploring nondeterministic causal models for better analysis of complex relationships.
― 6 min read
AGILE uses attention mechanisms to improve continual learning and reduce forgetting.
― 5 min read
A dataset examining how cultural values shape AI responses to social issues.
― 6 min read
This study questions the effectiveness of ReAct in enhancing LLM performance.
― 6 min read
Discover how AI agents learn to excel in Atari games using reinforcement learning.
― 6 min read
Examining the importance of data valuation for language models and its implications.
― 7 min read
A new method enhances fine-tuning of large models using spectral information.
― 5 min read
A new approach aligns language models with video content using textual simulations.
― 6 min read
A novel method combines LLMs and KGs for improved reasoning accuracy.
― 6 min read
Exploring new methods for enhancing time-series predictions using large datasets.
― 7 min read
Dynamic Model Predictive Shielding enhances safety and efficiency in reinforcement learning.
― 6 min read
A new method offers insights into the trustworthiness of LLM responses.
― 6 min read
Exploring how Federated Learning can enhance data privacy in healthcare.
― 5 min read
A new framework enhances language model learning through interactive feedback.
― 7 min read
A new method using reinforcement learning to improve energy trading strategies safely.
― 6 min read
GridTST enhances time series forecasting by integrating temporal and variate information.
― 7 min read
This paper presents new sampling methods for better generative model performance.
― 6 min read
A new scoring method improves the quality of synthetic anomalies in machine learning.
― 7 min read
Learn how to adjust weight decay for better model performance in AdamW.
― 7 min read
A novel approach improves efficiency in knowledge graph completion tasks.
― 5 min read
A new method for improving entity typing accuracy in knowledge graphs.
― 7 min read
Exploring new models for improved long-term predictions in various sectors.
― 7 min read
A new scheduler improves communication in Industrial Internet of Things using reinforcement learning.
― 5 min read
Automated tools improve code quality and streamline the review process.
― 9 min read
MotionCraft generates realistic videos using physics without extensive training.
― 8 min read
Integrating textual information enhances accuracy in time series forecasting.
― 5 min read
This article discusses the integration of AI into education and its implications.
― 8 min read
A new framework reduces data biases improving instruction-following performance in language models.
― 7 min read
A new audio codec offering high-quality compression and rich semantic content.
― 6 min read
NFCL combines accuracy and clarity in multivariate time-series predictions.
― 7 min read
New language models show promise in understanding and generating human language.
― 5 min read