AI Converts Brain Signals to HQ Images
Plus: Meta’s AI behind FB & Insta recommendations. OpenFlamingo V2 launched 5 new models.
Hello Engineering Leaders and AI Enthusiasts!
Welcome to the 53rd edition of The AI Edge newsletter. This edition brings you “AI Coverts Brain Signals to HQ Images.”
And a big thanks to all our incredible readers! 😊😊
In today’s edition:
🖼️ DreamDiffusion: Generating High-Quality Images from Brain EEG Signals
🔍 Meta discloses AI behind Facebook and Instagram recommendations
🌟 OpenFlamingo V2 launched 5 newly trained multimodal models
Let’s go!
DreamDiffusion: AI Generates High-Quality Images from Brain EEG Signals
New research has proposed DreamDiffusion, a novel method for generating high-quality images directly from brain electroencephalogram (EEG) signals, without the need to translate thoughts into text. It leverages powerful pre-trained text-to-image diffusion models to generate realistic images from EEG signals only, which is a non-invasive and easily obtainable source of brain activity.
Much recent AI research has attempted to reconstruct visual information based on fMRI (functional Magnetic Resonance Imaging) signals and has demonstrated the feasibility of reconstructing high-quality results from brain activities. However, they are still far from using brain signals to create conveniently and efficiently. DreamDiffusion addresses these issues.
Why does this matter?
The use of brain signals, including fMRI and EEG, to generate images has been an active area of AI research. This is a significant step towards portable and low-cost “thoughts-to-image'' AI, with potential applications in neuroscience, computer vision, and more.
Meta disclosed AI behind Facebook and Instagram recommendations
Meta is sharing 22 system cards that explain how AI-powered recommender systems work across Facebook and Instagram. These cards contain information and actionable insights everyone can use to understand and customize their specific AI-powered experiences in Meta’s products.
Moreover, Meta also shared its top ten most important prediction models rather than everything in the system to not dive into much technical detail can sometimes obfuscate transparency.
Why does this matter?
Reportedly, Meta is working on the next version of its open-source large-language model—technology that can power chatbots like ChatGPT—to make them available for commercial use. This can have big implications for other AI developers and businesses that are increasingly adopting it as well as individuals.
OpenFlamingo V2 launched 5 newly trained multimodal models
OpenFlamingo V2 has been unleashed with Five trained new models, spanning the 3B, 4B, and 9B scales, which have been introduced based on Mosaic's MPT-1B and 7B and Together.xyz's RedPajama-3B. These models are built on open-source models with less restrictive licenses than LLaMA.
OpenFlamingo models achieve over 80% of the performance of their Flamingo counterparts when looking at results from 7 evaluation datasets. Moreover, OpenFlamingo-3B and OpenFlamingo-9B achieve over 60% of the best fine-tuned performance by using only 32 in-context examples.
To further enhance the capabilities of OpenFlamingo, the training and evaluation code has undergone significant improvements. The evaluation suite now includes new datasets like TextVQA, VizWiz, HatefulMemes, and Flickr30k, expanding the scope of evaluation possibilities.
OpenFlamingo models can handle mixed sequences of images and text to generate text outputs. This enables the models to handle tasks like captioning, visual question answering, and image classification using relevant examples.
Why does this matter?
By incorporating images and text together, OpenFlamingo's V2 new models excel in tasks like image captioning, where descriptive text is generated based on the visual content.
This integration of visual and textual understanding not only enhances the performance of AI systems but also opens doors for advancements in content generation, human-computer interaction, and intelligent decision-making.
What Else Is Happening❗
🚀 CSM AI extraordinary release: Any Image to 3D! (Link)
💥 Meituan acquired Founder's recently launched 'OpenAI for China' for $234M. (Link)
🔥 Google announces the first Machine Unlearning challenge. (Link)
💰 Inflection secures $1.3B investment to build more ‘personal’ AI for everyone. (Link)
💡 Microsoft adds new AI Shopping Tools to Bing & Edge. (Link)
💎 Runway AI valued at $1.5 billion in latest funding. (Link)
✨ Microsoft empowers Moody's with Gen AI integration! (Link)
🛠️ Trending Tools
Weedone: Task limits, weekly goals, and efficiency are crucial. AI copilot generates weekly goals and OKRs.
Formx AI: Train ML-powered no-code extractors. Automate data extraction from documents with 90%+ accuracy via API integration.
Junia: Generate AI-powered 6000-word articles, images, and SEO metadata within minutes. Chinese version available on heyjunia.cn.
Signum AI: Track contacts' behavior using public data from your CRM. Determine the ideal moment for reconnecting.
BlogSEO AI: Create high-quality, user-first, and search-engine-friendly blog articles in minutes with AI content generation.
Summer AI: Narrates points of interest, events. AR mode available.
Stork: AI-powered video-conferencing. Capture, summarize calls. Share recordings.
GM Plus: Boost email writing efficiency with AI-generated prompts, templates.
That's all for now!
If you are new to ‘The AI Edge’ newsletter. Subscribe to receive the ‘Ultimate AI tools and ChatGPT Prompt guide’ specifically designed for Engineering Leaders and AI enthusiasts.
Thanks for reading, and see you tomorrow. 😊