A comprehensive index of artificial intelligence and machine-learning research with AI-generated summaries, citation metrics, and direct links to papers and code.
Michael Dowling, Brian M. Lucey
ChatGPT can assist finance research by generating ideas and identifying data, but is weaker on literature synthesis and testing frameworks, with output quality depending on private data and domain expertise.
Jürgen Schmidhuber, Daan Wierstra, Matteo Gagliolo, et al.
Evolino combines evolution of recurrent neural network hidden weights with linear output weight computation, enabling LSTM to solve tasks where gradient methods fail.
Norman P. Jouppi, Cliff Young, Nishant Patil, et al.
Evaluates a custom ASIC (TPU) for neural network inference, showing 15-30x speedup and 30-80x TOPS/Watt over contemporary CPUs and GPUs.
Tomohiro Takagi, Michio Sugeno
This paper introduces a fuzzy identification method using fuzzy implications and linear input-output relations for modeling and control of systems.
J. Kirchner, Yining Chen, Harri Edwards, et al.
Proposes legibility training via a Prover-Verifier Game to make LLM chain-of-thought reasoning easier for humans to verify, improving trust in model outputs.
Unknown
QALIGN uses MCMC sampling and Minimum Bayes Risk to align language model outputs at test time without retraining.
Unknown
ShieldGemma is a suite of LLM-based safety content moderation models built on Gemma2 that predict safety risks across key harm types in user input and LLM-generated output.
Unknown
CriticGPT uses RLHF to train a GPT-4-based model that critiques ChatGPT code outputs, helping humans catch bugs more accurately.
Unknown
ProRL shows that prolonged RL training with KL control and policy resetting can discover novel reasoning strategies beyond base model outputs.
Unknown
A multimodal transformer model pre-trained to predict the next token, accepting image and text inputs and producing text outputs.
Unknown
PaLM is a 540-billion parameter densely activated Transformer model trained efficiently across multiple TPU Pods using the Pathways system.
Unknown
Introduces speculative decoding, an algorithm for faster sampling from autoregressive models by computing multiple tokens in parallel without altering outputs.