Big Model Daily on December 8

News6months agorelease AIWindVane
26 0

Big Model Daily on December 8[Big Model Daily on December 8] He Kaiming’s new work: A simple framework achieves new SOTA for unconditional image generation! Completed in cooperation with MIT; Gemini’s first day online: Users had mixed reviews, the demo was questioned as “fake”, and Google admitted it; OpenAI employees revealed that they did not want to go to Microsoft at all, and joint resignation was only a last resort; Musk’s Grok large model can be played Got it

To bridge the gap between chemical reaction pre-training and conditional molecule generation, Peking University & Wangshi Intelligence propose a “unified” model

Recently, a research team from Peking University and Wangshi Intelligence proposed a new method to bridge the gap between reaction-based molecular pre-training and generation tasks. Inspired by the mechanisms of organic chemistry, the researchers developed a new pre-training framework that enables it to incorporate inductive bias into models. The proposed framework achieves state-of-the-art results in performing challenging downstream tasks. By leveraging knowledge of chemistry, generative frameworks overcome the limitations of current molecular generation models that rely on a small number of reaction templates. In extensive experiments, the model generated high-quality synthesizable drug-like structures. Overall, this research represents an important step toward a large-scale deep learning framework for a variety of React-based applications.

OpenAI employees revealed: They don’t want to go to Microsoft at all, and joint resignation to force the palace is only a last resort

OpenAI’s infighting is about to start its second season.

Helen Toner, the former board member who was “kicked out”, released the latest inside story: First, Altman was fired to strengthen OpenAI; second, her resignation was coerced by the company’s lawyers.

AMD’s next GPU is a 3D integrated superchip: MI300 combines 13 pieces of silicon into one chip

AMD unveiled its next-generation AI accelerator chip, the Instinct MI300, at its recent AMD Advancing AI event, an unprecedented feat of 3D integration. The MI300 will power the El Capitan supercomputer, a layer cake of compute, memory and communications that is three silicon wafers high and can transfer up to 17 terabytes of data vertically between those silicon planes. It can make certain machine learning-critical calculations 3.4x faster. The chip is both different and similar to competing products such as Nvidia’s Grace-Hopper superchip and Intel’s supercomputer accelerator Ponte Vecchio.

Musk’s Grok large model is now playable!

Musk suddenly announced that the Grok large model was open to paying users in large quantities. Click on it and find that Grok has turned on the weird and fun mode by default. Let it tell an Elon-ish joke (give me an elon-ish joke), he said: Why Elon Musk brought a ladder to the math test because he wanted to Elon-gate (stretch, lengthen) the results.

Alibaba Intelligent Information Business Group: Quark’s large model has passed the filing

Alibaba Intelligent Information Business Group Office learned that its self-developed Quark large model has passed the filing. It will continue to upgrade content products and intelligent tools in the fields of general education, health, and creation, and launch a series of AIGC (generative artificial intelligence) innovative applications. It will comprehensively upgrade Quark with the help of large model capabilities to improve users’ learning, work, and life. s efficiency.

It was revealed that the person in charge of the Dojo supercomputing project has changed, which may affect Tesla’s autonomous driving technology

According to people familiar with the matter, Ganesh Venkataramanan, the project leader of Tesla’s Dojo supercomputer, resigned in November. Media analysts believe that this resignation not only dealt a blow to this expensive project, but also caused Tesla’s autonomous driving technology to encounter new setbacks.

It is rumored that Apple will make a major upgrade to the microphone. Could iPhone 16 be the company’s first AI phone?

Well-known fruit chain analyst Ming-Chi Kuo pointed out that given the popularity of generative artificial intelligence in 2023, it is not difficult to predict that Apple will launch related innovative features in iPhone 16; voice input will be a key interface for AI, so Siri’s The functions and specifications of software and hardware are the key to promoting AIGC.

Cisco launches Cisco AI Assistant security assistant that can intelligently set up firewall rules to prevent hacker attacks

According to Cisco’s official press release, Cisco recently launched a security tool, Cisco AI Assistant, which it claims can help users manage firewall rules and prevent hackers from taking advantage of the situation.

© Copyright notes

Related posts

No comments

No comments...