A comprehensive index of artificial intelligence and machine-learning research with AI-generated summaries, citation metrics, and direct links to papers and code.
Davy Tsz Kit Ng, Jac Ka Lok Leung, Samuel Kai Wah Chu, et al.
This exploratory review conceptualizes AI literacy by proposing four aspects—know and understand, use and apply, evaluate and create, and ethical issues—based on 30 peer-reviewed articles.
Jürgen Rudolph, Samson Tan, Shannon Tan
An early peer-reviewed exploration of ChatGPT's implications for higher education, focusing on assessment, learning, and teaching, with recommendations for stakeholders.
Unknown
RidgeloRA enhances LoRA by incorporating matrix ridge regression to improve parameter-efficient fine-tuning of large language models.
Unknown
An in-depth exploration of factors affecting multi-modal in-context learning, using semantic diversity analysis to understand performance drivers.
Unknown
LoRA-XS introduces a low-rank adaptation method that drastically reduces trainable parameters while maintaining performance.
Unknown
The paper introduces nested low-rank adaptation, a method that reduces the number of trainable parameters for fine-tuning large models compared to standard LoRA.
Unknown
Relax introduces latent exploration for RLVR to improve reasoning in large reasoning models.
Unknown
ETO is a novel learning approach that uses exploration-based trajectory optimization to improve LLM agents through trial and error.
Guanghao Ye, Khiem Pham, Xinzhi Zhang, et al.
Proposes RLSP, a post-training framework that decouples exploration and correctness signals during PPO to enable emergent reasoning behaviors in LLMs.
Qi Sun, Edoardo Cetin, Yujin Tang
Transformer-Squared is a self-adaptive LLM framework that adjusts only the singular components of weight matrices in real-time, using a two-pass dispatch and reinforcement-learned expert vectors to outperform LoRA with fewer parameters.
Rujikorn Charakorn, Edoardo Cetin, Shinnosuke Uesaka, et al.
Doc-to-LoRA (D2L) is a lightweight hypernetwork that meta-learns to perform approximate context distillation in a single forward pass, generating LoRA adapters for a target LLM to reduce inference latency and KV-cache memory.
Roeland Scheepens, Niels Willems, Huub van de Wetering, et al.
This paper introduces a flexible architecture for multivariate trajectory density maps using composable blocks and expressions, enabling analysts to model domain knowledge for visual exploration.