Joshua Pritsker explores optimization techniques to allow for full parameter 16-bit finetuning of language models up to 7B on a single 24GB GPU

Taylor Webb, Shanka Subhra Mondal, Chi Wang, Brian Krabach, and Ida Momennejad propose a black box architecture with multiple LLM-based (GPT-4) modules to improve planning in large language models. The architecture utilizes specialized prefrontal cortex-inspired modules for conflict monitoring, state prediction, state evaluat...

Timothée Darcet, Maxime Oquab, Julien Mairal and Piotr Bojanowski propose a simple yet effective solution based on providing additional tokens to the input sequence of the Vision Transformer for internal computation

Hu Xu, Saining Xie, Xiaoqing Ellen Tan, Po-Yao Huang, Russell Howes, Vasu Sharma, Shang-Wen Li, Gargi Ghosh, Luke Zettlemoyer and Christoph Feichtenhofer demystify the data used in Contrastive Language-Image Pre-training (CLIP) and introduce Metadata-Curated Language-Image Pre-training (MetaCLIP) to improve CLIP's data curati...

Meta present a series of long-context LLMs that support effective context windows of up to 32,768 tokens

Quantamagazine - Under a microscope, cells in a worm embryo eliminated one-third of their genome — an uncompromising tactic that may combat genetic parasites.

Economic Innovation Group release a policy brief supporting "Chipmaker's" visas to attract skilled workers in context of the CHIPS Act.

OpenAi publish the system card of their multimodal model GPT4V, outlining how they prepared the vision capabilities of GPT-4 for deployment.

Microsoft Research describe a framework for simplifying the orchestration, optimization, and automation of LLM workflows.

R. Thomas McCoy, Shunyu Yao, Dan Friedman, Matthew Hardy and Thomas L. Griffiths evaluate LLMs based on probability of the task to be performed, the probability of the target output, and the probability of the provided input.

Andreas Tsamados, Luciano Floridi and Mariarosaria Taddeo offer recommendations designed to mitigate critical vulnerabilities introduced by autoregressive-large language models.

VIDEO | Jeremy Howard (fast.ai) takes you on a comprehensive journey through the fascinating landscape of LMs.

Stable Diffusion Art propose a comprehensive guide to use Stable Diffusion and ControlNet.

Semianalysis report on the acquisition of Illumina by Nvidia, and use the opportunity to go through all the opportunities of AI in healthcare

Daniel Miessler shares his thoughts on the impact of AI on the economy, in particular on downside impact for B2B services

NewScientist - Surgeons have successfully transplanted a genetically modified pig heart into a human for the second time.

The CDC 2022 Adult Obesity Prevalence Maps show the proportion of US adults with a body mass index (BMI) equal to or greater than 30 ( ≥30 kg/m2) based on self-reported weight and height.

Yukang Chen, Shengju Qian, Haotian Tang, Xin Lai, Zhijian Liu, Song Han and Jiaya Jia speed up the context extension of LLMs by revisiting the parameter-efficient fine-tuning regime for context expansion.

Quantamagazine - Some modern image generators rely on the principles of diffusion to create images. Alternatives based on the process behind the distribution of charged particles may yield even better results.

The European Supervisory Authorities (ESAs), together with the competent authorities (CAs), in 2022 carried out a joint high-level analysis with the objective of getting a preliminary overview of the provision of information and communication technology (ICT) services to the EU financial entities by ICT third-party service p...

In this paper, Vincent Perot, Kai Kang, Florian Luisier, Guolong Su, Xiaoyu Sun, Ramya Sree Boppana, Zilong Wang, Jiaqi Mu, Hao Zhang amd Nan Hua introduce Language Model-based Document Information Extraction and Localization (LMDX), a methodology to adapt arbitrary LLMs for document information extraction from OCR'ed documen...

Fabrizio Dell'Acqua, Edward McFowland, Ethan Mollick, Hila Lifshitz-Assaf, Katherine Kellogg, Saran Rajendran, Lisa Krayer, François Candelon, Karim R Lakhani conducted a study with Boston Consulting Group to examine the performance implications of AI on realistic, complex, and knowledge-intensive tasks.

Paul Vanags, Jo Cutler, Patricia Lockwood and Fabian Kosse use data from 80,337 people across 76 countries to show that both objective wealth (household income), and subjective wealth (financial well-being), are positively and consistently associated with higher prosociality.

Elliot Hershberg explores the story of Illumina, a company that established a Measurement Monopoly in the life sciences by inventing a profound new measurement technology and pairing it with a successful business strategy. Founded by a blind venture capitalist, a veterinarian, a Harvard professor, a chemist, and a molecular b...

Jon Saad-Falcon, Joe Barrow, Alexa Siu, Ani Nenkova, Ryan A. Rossi and Franck Dernoncourt propose an approach called PDFTriage that enables models to retrieve the context based on either structure or content.