stp2y - CybAI news

25 May

Research Papers in Oct 2023: A Potential Successor to RLHF for Efficient LLM Alignment and the Resurgence of CNNs

stp2y0 CommentsAI

From Vision Transformers to innovative large language model finetuning techniques, the AI community has been very active with lots of interesting research this past month.Here's a snapshot of the highlights I am covering in this article:In the paper ConvNets Match Vision Transformers at Scale, Smith et al. invest significant computational resources to conduct a thorough comparison between Convolutional Neural Networks (CNNs) and Vision Transformers (ViTs), challenging the prevailing notion that ViTs outperform CNNs in image classification tasks. The Mistral 7B paper introduces a compact yet powerful language model that, despite its relatively modest size of 7 billion tokens, outperforms its larger…

25 May

If Scarlett Johansson can’t bring the AI firms to heel, what hope for the rest of us? | John Naughton

stp2y0 CommentsChat-GPT

On Monday 13 May, OpenAI livestreamed an event to launch a fancy new product – a large language model (LLM) dubbed GPT-4o – that the company’s chief technology officer, Mira Murati, claimed to be more user-friendly and faster than boring ol’ ChatGPT. It was also more versatile, and multimodal, which is tech-speak for being able to interact in voice, text and vision. Key features of the new model, we were told, were that you could interrupt it in mid-sentence, that it had very low latency (delay in responding) and that it was sensitive to the user’s emotions.Viewers were then treated…

25 May

ArXiv Pre-Print “Evaluating AI Systems under Uncertain Ground Truth: a Case Study in Dermatology” • David Stutz

stp2y0 CommentsGenAI

ArXiv Pre-Print “Evaluating AI Systems under Uncertain Ground Truth: a Case Study in Dermatology” In supervised machine learning, we usually assume access to ground truth label for evaluation. In many applications, however, these ground truth labels are derived from expert opinions. Disagreement among these experts is typically ignored using simple majority voting or averaging. Unfortunately, this can have severe consequences by over-estimating performance or mis-guiding model selection. In our work presented in this article, we tackle this problem by introducing a statistical framework for aggregating expert opinions. Abstract For safety, AI systems in health undergo thorough evaluations before deployment, validating…

25 May

Data Machina #246

stp2y0 CommentsViral News

New Trends in Vision-Language Models (VLMs.) The evolution of VLMs in recent months has been pretty impressive. Today VLMs exhibit some amazing capabilities. See the two links below on what VLMs can do and how they work:But still VLMs are facing some challenges for example in terms of: multimodal training datasets, resolution, long-form modality, vision-language integration, and concept understanding. Somewhat along those lines, I see 5 trends happening in VLMs: 1) VLMs run on local environment 2) Emerging VLM videoagents 3) Unified structure learning for VLMs 4) Personalisation of VLMs and 5) Fixing the VLM resolution curse. Let’s see…VLMs on…

25 May

The AI Revolution Will Not Be Monopolized: Behind the scenes

stp2y0 CommentsNews

FabNER Claude 2 accuracy on # of examples 10 20 30 40 50 60 70 80 90 100 0 100 200 300 400 500 20 examples F-Score Speed (words/s) GPT-3.5 1 78.6 < 100 GPT-4 1 83.5 < 100 spaCy 91.6 4,000 Flair 93.1 1,000 SOTA 2023 2 94.6 1,000 SOTA 2003 3 88.8 > 20,000 1. Ashok and Lipton (2023), 2. Wang et al. (2021), 3. Florian et al. (2003) SOTA on few- shot prompting RoBERTa-base * * EXPERIMENTS * * * EXPERIMENTS * * * EXPERIMENTS * * * CoNLL 2003: Named Entity Recognition Source link lol

25 May

Viruses are doing mysterious things everywhere – AI can help researchers understand what they’re up to in the oceans and in your gut

stp2y0 CommentsChat-GPT

Viruses are a mysterious and poorly understood force in microbial ecosystems. Researchers know they can infect, kill and manipulate human and bacterial cells in nearly every environment, from the oceans to your gut. But scientists don’t yet have a full picture of how viruses affect their surrounding environments in large part because of their extraordinary diversity and ability to rapidly evolve. Communities of microbes are difficult to study in a laboratory setting. Many microbes are challenging to cultivate, and their natural environment has many more features influencing their success or failure than scientists can replicate in a lab. So systems…

25 May

PeerDB raises $3.6M to accelerate PostgreSQL data movement – SiliconANGLE

stp2y0 CommentsGenAIMike Wheatley, PeerDB raises $3.6M to accelerate PostgreSQL data movement, SiliconANGLE

Startup PeerDB Inc., which has built a data movement platform specifically for open-source PostgreSQL database management systems, said today it has closed on a seed funding round worth $3.6 million. Leading the investment was 8VC, and it was joined by Y Combinator, Wayfinder Ventures, Webb Investment Network, Flex Capital, Rogue Capital, Pioneer Fund, Orange Collective and several angel investors. PeerDB says its data movement platform is designed to improve on the capabilities of existing data pipelines that were never designed for Postgres databases. Existing data movement and extract, transform and load or ETL tools often prioritize the sheer number of connectors…

25 May

Coding With Devin: My New AI Programming Agent

stp2y0 CommentsRAG models

Sponsored By: Reflect This essay is brought to you by Reflect, an ultra-fast notes app with an AI assistant built in directly. Simplify your note-taking with Reflect's advanced features, like custom prompts, voice transcription, and the ability to chat with your notes effortlessly. Elevate your productivity and organization with Reflect. Was this newsletter forwarded to you? Sign up to get it in your inbox.We onboarded four new engineers at Every this week. The onboarding process was what you'd expect. I gave them access to our GitHub account so they could download our main code repos. I was peppered with the usual…

25 May

AI vs. Humanity: Who Will Come Out on Top?

stp2y0 CommentsViral News

The below is a summary of my recent article on superintelligence. Elon Musk predicts that Artificial Superintelligence (ASI) will emerge by 2025, much earlier than his previous estimates. While Musk's track record with predictions is mixed, this one sparks serious contemplation about the future. The moment AI surpasses human cognitive abilities, known as the singularity, will usher in a new era with both unprecedented possibilities and profound perils. As we edge closer to this event horizon, it's essential to ask if we are prepared to navigate the uncertainties and harness the potential of AI responsibly. The journey towards ASI has…

25 May

Generating 3D Molecular Conformers via Equivariant Coarse-Graining and Aggregated Attention

stp2y0 CommentsAIcoarse, conditional generation, equivariance, geometric deep learning, gnn, hierachical, molecule

Figure 1: CoarsenConf architecture. Molecular conformer generation is a fundamental task in computational chemistry. The objective is to predict stable low-energy 3D molecular structures, known as conformers, given the 2D molecule. Accurate molecular conformations are crucial for various applications that depend on precise spatial and geometric qualities, including drug discovery and protein docking. We introduce CoarsenConf, an SE(3)-equivariant hierarchical variational autoencoder (VAE) that pools information from fine-grain atomic coordinates to a coarse-grain subgraph level representation for efficient autoregressive conformer generation. (more…)