- AI Research Insights
- Posts
- AI News: 🚀 Stability AI releases DeepFloyd IF | Microsoft AI Team Unveils NaturalSpeech 2 | Hugging Face Launches HuggingChat.....
AI News: 🚀 Stability AI releases DeepFloyd IF | Microsoft AI Team Unveils NaturalSpeech 2 | Hugging Face Launches HuggingChat.....
This newsletter brings AI research news that is much more technical than most resources but still digestible and applicable
➡️ Stability AI releases DeepFloyd IF, a powerful text-to-image model that can smartly integrate text into images. DeepFloyd IF is a state-of-the-art text-to-image model released on a non-commercial, research-permissible license that provides an opportunity for research labs to examine and experiment with advanced text-to-image generation approaches. In line with other Stability AI models, Stability AI intends to release a DeepFloyd IF model fully open source at a future date.
➡️ This paper introduces JaxPruner: an open-source JAX-based pruning and sparse training library for machine learning research. JaxPruner aims to accelerate research on sparse neural networks by providing concise implementations of popular pruning and sparse training algorithms with minimal memory and latency overhead. Algorithms implemented in JaxPruner use a common API and work seamlessly with the popular optimization library Optax, which, in turn, enables easy integration with existing JAX based libraries.
➡️ Prompting Unlocks Theory-of-Mind in LLMs: Large language models (LLMs) excel in many tasks in 2023, but they still face challenges in complex reasoning. Theory-of-mind (ToM) tasks, which require understanding agents' beliefs, goals, and mental states, are essential for common-sense reasoning involving humans, making it crucial to enhance LLM performance in this area. This study measures the ToM performance of GPT-4 and three GPT-3.5 variants (Davinci-2, Davinci-3, GPT-3.5-Turbo), and investigates the effectiveness of in-context learning in improving their ToM comprehension.
➡️ Hugging Face Launches HuggingChat: An Open-Source Alternative To ChatGPT. HuggingChat, like ChatGPT, is versatile and can do things like write code, write emails, and write rap lyrics. Open Assistant, a project managed by the German non-profit LAION, created the AI model that powers HuggingChat. To generate the assistant of the future, Open Assistant aspires to mimic ChatGPT’s features while also including APIs, doing on-the-fly research, and more.
➡️ Meet Segment AnyRGBD: A Toolbox To Segment Rendered Depth Images Based On SAM: To segment rendered depth pictures using SAM; researchers have developed the Segment AnyRGBD toolkit. SAD, short for Segment Any RGBD, was recently introduced by NTU researchers. SAD can easily segment any 3D object from RGBD inputs (or generated depth pictures alone).
➡️ No time to get ready for your meeting? Use xpression camera to be dressed for all occasions. The Xpression Camera is equipped with state-of-the-art AI algorithms that analyze your appearance and surroundings to create realistic, high-quality digital clothing overlays. These overlays can be customized to suit your preferences and seamlessly blend with your actual clothing, creating a convincing and stylish look.
➡️ Microsoft AI Team Unveils NaturalSpeech 2: A Cutting-Edge TTS System with Latent Diffusion Models for Powerful Zero-Shot Voice Synthesis and Enhanced Expressive Prosodies. The researchers began by training a neural audio codec that uses a codec encoder to transform a speech waveform into a series of latent vectors and a codec decoder to restore the original waveform. After obtaining previous vectors from a phoneme encoder, a duration predictor, and a pitch predictor, they use a diffusion model to construct these latent vectors.
➡️ Can an artificial intelligence chatbot assistant respond to patient questions of comparable quality and empathy to those written by physicians? A new breakout study compared ChatGPT vs. physician responses to patient questions. Overall, patients preferred the chatbot over the physician 79% of the time. Notably, 4 questions received a 100% preference for the chatbot over the physician. It's clear to see the overall quality and empathy improvement in ChatGPT's answer.
AI Tools Club: Find 100s of cool artificial intelligence (AI) tools. Our expert team reviews and provides insights into some of the most cutting-edge AI tools available.
Do You Know Marktechpost has a community of 1.5 Million+ AI Professionals and Engineers? For partnership and advertisement, please feel free to contact us through this form.