- NLPlanet Newsletter
- Posts
- Weekly AI and NLP News — May 9th 2023
Weekly AI and NLP News — May 9th 2023
New open-source LLMs, a code-generating model from HF, and works on search engines
Here are your weekly articles, guides, and news about NLP and AI chosen for you by NLPlanet!
😎 News From The Web
Google Plans to Make Search More ‘Personal’ with AI Chat and Video Clips. Google is shifting the way it presents search results to incorporate conversations with artificial intelligence, along with more short video and social media posts, a departure from the list of website results that has made it the dominant search engine for decades.
Introducing MPT-7B: A New Standard for Open-Source, Commercially Usable LLMs, by MosaicML. MPT-7B is a transformer trained from scratch on 1T tokens of text and code. It is open-source, available for commercial use, and matches the quality of LLaMA-7B.
Releasing 3B and 7B RedPajama-INCITE family of models including base, instruction-tuned & chat models. The release includes the first models trained on the RedPajama base dataset: a 3 billion and a 7B parameter base model that aims to replicate the LLaMA recipe as closely as possible.
Hugging Face and ServiceNow release a free code-generating model. AI startup Hugging Face and ServiceNow Research have released StarCoder, a free alternative to code-generating AI systems along the lines of GitHub’s Copilot.
Microsoft’s Bing Chat AI is now open to everyone, with plug-ins coming soon. All you need is a Microsoft account to get access to the GPT-4-powered version of Bing.
Hackers are increasingly using ChatGPT lures to spread malware on Facebook. As public interest in generative AI chatbots grows, hackers are increasingly using ChatGPT-themed lures to spread malware across Facebook, Instagram, and WhatsApp.
The amazing AI super tutor for students and teachers, Khan Academy. Sal Khan, the founder and CEO of Khan Academy, thinks artificial intelligence could spark the greatest positive transformation education has ever seen. He demos some exciting new features for their educational chatbot, Khanmigo.
“Godfather of AI” quits Google with regrets and fears about his life’s work. Geoffrey Hinton who won the ‘Nobel Prize of computing’ for his trailblazing work on neural networks is now free to speak about the risks of AI.
Scientists Use GPT AI to Passively Read People's Thoughts in Breakthrough. An AI model similar to ChatGPT was combined with fMRI readings to non-invasively decode continuous language from subjects, a new study reports.
📚 Guides From The Web
Let's talk about ChatGPT with Code Interpreter & Microsoft Copilot. Between the expanding capabilities of GPT-4, and the soon-to-be everywhere Microsoft Copilot, work is going to start changing in a matter of months, not years.
Text-to-Video: The Task, Challenges and the Current State. This blog post discusses the past, present, and future of text-to-video models. It starts by reviewing the differences between the text-to-video and text-to-image tasks, and discusses the unique challenges of unconditional and text-conditioned video generation.
Google "We Have No Moat, And Neither Does OpenAI". A Leaked internal Google document claims open-source AI could outcompete Google and OpenAI. The opinions are just from a Google employee, not from the firm.
🔬 Interesting Papers and Repositories
mlc-ai/mlc-llm: Enable everyone to develop, optimize and deploy Large Language Models natively on everyone's devices.
Significant-Gravitas/Auto-GPT-Plugins: Plugins for Auto-GPT.
Distilling Step-by-Step! Outperforming Larger Language Models with Less Training Data and Smaller Model Sizes. The authors introduce Distilling step-by-step, a new mechanism that trains smaller models that outperform LLMs, and achieves so by leveraging less training data needed by finetuning or distillation.
Unlimiformer: Long-Range Transformers with Unlimited Length Input. A general approach that can wrap any existing pretrained encoder-decoder transformer, and offload the attention computation across all layers to a single k-nearest-neighbor index.
Thank you for reading! If you want to learn more about NLP, remember to follow NLPlanet. You can find us on LinkedIn, Twitter, Medium, and our Discord server!