AI Decodes Odors Like Humans 🌿
PLUS: Generative AI Courses on Google Cloud, LLMs for Solving P vs NP Problems
Today’s top AI Highlights:
Neural Networks Map Odors Like Humans
Einstein Copilot for Salesforce
Generative AI Courses on Google Cloud
Using LLMs for Solving P vs NP Problems
& so much more!
Read time: 3 mins
Latest Developments 🌍
Neural Networks for Odor Perceptions 💡👃
In order to understand how molecules relate to our sense of smell, researchers have used advanced neural networks to create a map that predicts odors like a human.
Key Highlights:
A dataset of around 5000 molecules, each described by multiple odor labels, is used to train the model, achieving strong cross-validation performance.
Graph neural networks (GNN) are employed to create a Principal Odor Map (POM) that accurately preserves perceptual relationships and predicts odor quality for new odorants.
The model's odor description performance matches that of a human, and surpasses chemoinformatic models in various odor prediction tasks.
Einstein Copilot for Next-Gen CRM 🫂
Salesforce has unveiled the next generation of Einstein, introducing Einstein Copilot, a powerful conversational AI assistant, along with Einstein Copilot Studio.
Key Highlights:
Einstein Copilot enhances user productivity, providing natural language support and secure access to proprietary company data within every Salesforce application.
Einstein Copilot Studio can be used by businesses to create custom AI-powered apps, prompts, and actions to streamline operations, boost sales, and improve customer service across various channels.
Both Einstein Copilot and Studio operate within the secure Einstein Trust Layer, preserving data privacy while delivering high-quality AI results.
New Generative AI Courses on Google Cloud🔖
Google Cloud is releasing comprehensive Generative AI training resources, offering both introductory and advanced content to explore the technology.
Key Highlights:
Google Cloud provides two distinct learning paths, catering to non-technical and technical audiences, allowing learners to delve into Generative AI at their own pace.
Individuals with no technical background can kickstart their journey with introductory courses, earning valuable skill badges along the way.
For those seeking hands-on experience, Google Cloud offers a range of labs covering various aspects of Generative AI, enabling users to gain practical expertise in the field.
LLMs and Socratic Reasoning for P vs NP Problems 🧩
The P vs. NP problem involves classifying computational problems based on their solvability and has significant implications for cryptography and optimization. It has remained unsolved for over fifty years. Researchers in a study have investigated how LLMs can be used to accelerate progress on the P versus NP problem.
Key Highlights:
The research introduces the Socratic Reasoning framework, inspired by Socrates, enabling LLMs to engage in critical thinking and complex problem-solving.
GPT-4 collaborates with human experts to tackle the elusive problem, demonstrating its capability to explore highly complex issues.
This innovative approach highlights the potential of LLMs to discover novel insights within the extensive solution space, offering new prospects for scientific discoveries in the realm of LLM for Science.
Tools of the Trade ⚒️
Deekard: Analyse your database swiftly with questions in natural language, create dashboards in minutes, along with AI-powered SQL editor.
Parea AI: Create LLM-powered products and monitor their performance, manage and test prompts, and improve AI development workflows.
Bearing AI: AI-powered solutions for green shipping, simplifying compliance with the Carbon Intensity Indicator (CII), optimizing vessel performance, and more.
Owny: AI-powered platform for startups and funds to manage capital, streamline processes, and offers various financial tools and templates for efficient operations.
DealPage: AI-powered workflows for sales engineers, streamlining tasks such as responding to RFPs, managing knowledge bases, and handling technical queries
😍 Enjoying so far, TWEET NOW to share with your friends!
Hot Takes 🔥
thinking about how, in the last year, > 5 ML engineers have told me, unprompted, that they want to do less ML & more software engineering. not because it’s more lucrative to build ML platforms & devtools, but because models can be too unpredictable & make for a stressful job ~ Shreya Shankar
the thing that gives me most confidence that technological progress is divine providence is that it’s a force for good even in the most degenerate cases. somehow even the advancement of weapons technology has led to a pax atomica and unprecedented safety and growth ~ roon
AI has lots of limits, and it isn’t hard to find problems LLMs can’t solve. I would be much more concerned if there were no weaknesses. ~ Ethan Mollick
Meme of the Day 🤡
That’s all for today!
See you tomorrow with more such AI-filled content. Don’t forget to subscribe and give your feedback below 👇
Real-time AI Updates 🚨
⚡️ Follow me on Twitter @Saboo_Shubham for lightning-fast AI updates and never miss what’s trending!!
PS: I curate this AI newsletter every day for FREE, your support is what keeps me going. If you find value in what you read, share it with your friends by clicking the share button below!