Could there be anything accounting for general exposure to images of people in daily life? E.g. magazines for women depicting women, magazines for men depicting women. Or forward facing images with eye contact?
AI
-
GPT-4 now analyzes PDFs
By
–
GPT-4 can now analyze PDFs A moment of silence for the 10,000 AI startups that just became obsolete.
-
Confidential Computing Enhances Data Security During Processing
By
–
Confidential Computing enhances security by encrypting data during processing. While traditional methods encrypt data at rest and in transit, Confidential Computing ensures that data remains encrypted during computation using RAM encryption and more. Microblog @antgrasso
-
MimicGen: Scalable Robot Learning from Human Demonstrations
By
–
MimicGen: A Data Generation System for Scalable Robot Learning using Human Demonstrations Mandlekar et al.: https://
arxiv.org/abs/2310.17596 #Robotics #DeepLearning #ArtificialIntelligence -
Branch-Solve-Merge: Advanced LLM Reasoning Framework
By
–
10/ Branch-Solve-Merge Reasoning in LLMs – an LLM program that enables an LLM to plan a decomposition of a task into multiple parallel sub-tasks, independently solve them, and fuse solutions to the sub-tasks.
-
CommonCanvas: CC Dataset for Competitive Diffusion Model Training
By
–
8/ CommonCanvas – a dataset of Creative-Commons-licensed (CC) images to train diffusion models competitive with Stable Diffusion 2 (SD2); rely on a transfer learning technique to produce high-quality synthetic captions paired with curated CC images.
-
Managing AI Risks: Social Harms and Autonomous System Challenges
By
–
9/ Managing AI Risks – a short paper outlining risks from upcoming and advanced AI systems, including an examination of social harms, malicious uses, and other potential societal issues emerging from the rapid adoption of autonomous AI systems.
-
ConvNets Match Vision Transformers: Scaling Laws Analysis
By
–
7/ ConvNets Match Vision Transformers – evaluates a performant ConvNet architecture pretrained on JFT-4B at scale; observes a log-log scaling law between the held out loss and compute budget.
-
Detecting Pretraining Data from Large Language Models
By
–
6/ Detecting Pretraining Data from LLMs – explores the problem of pretraining data detection which aims to determine if a black box model was trained on a given text.
-
Matryoshka Diffusion Models for High-Resolution Image Video Synthesis
By
–
3/ Matryoshka Diffusion Models – introduces an end-to-end framework for high-resolution image and video synthesis; involves a diffusion process that denoises inputs at multiple resolutions jointly and uses a NestedUNet architecture…