AI News

Instruction fine-tuning approaches — fine-tuning large language models on tasks described via instructions — have shown promising results in improving zero- and few-shot learning performance through fine-tuning objectives, task sampling strategies, instruction-tuning benchmarks, training datasets and so on. Yet there is limited understanding with regard to the performance trade-offs involved in the various choices made during the instruction-tuning process

State space models (SSMs) designed for modelling dynamic systems have achieved outstanding sequence-to-sequence performance in fields ranging from time series analysis to audio generation. SSMs however struggle on language modelling tasks, where they cannot match the performance of transformer architectures. In the new paper Hungry Hungry Hippos: Towards Language Modeling with State Space Models, a research team from Stanford University and the State Unive

Stable Diffusion 2.0 model recipe: Out-of-the-box configuration for Stable Diffusion 2.0 in Colossal-AI enables low-cost training, fine-tuning, and inference, while also reducing GPU memory consumption by up to 5.6 times and hardware costs by up to 46 times. All of which can be achieved with just one line of code. BLOOM model recipe: The ready-made inclusion of the 175 billion parameter BLOOM model offers stand-alone inference with a 4-fold reduction in GP

Baidu, Inc. ,a leading AI company with strong internet foundation, today hosted its annual flagship developer conference Baidu Create 2022. The event was held virtually in XiRang, Baidu’s metaverse app, as the company’s first conference co-hosted by humans and robots. In addition to featuring several key technology advancements, Baidu Create offered an in-depth exploration of Baidu’s research and analysis of future technology trends, covering a range of em

If anonymously sourced reports are any indication, the long-delayed release of the new Siri, Apple’s AI-powered assistant updated for the supposedly AI-centric consumer expectations of the 2020s, is finally imminent. But a new report from Bloomberg’s Mark Gurman suggests that Apple is acting a little sheepish about some privacy compromises it may or may not have recently made as it has prepared Siri for prime time. To back up a bit, back in January, Google

After a grueling three-week trial that dragged some of the biggest names in the tech industry onto the witness stand to testify, made private correspondence and diaries public, and placed the details of private lives under scrutiny, the jury in Elon Musk’s case against OpenAI over its decision to become a for-profit entity came to a decision: it is barred by the statute of limitations and no decision on the legitimacy of the claims will be provided. On the

Elon Musk’s xAI recently offered employees payments of $420 to hand over their completed tax returns for use in training Grok, according to a new report from Bloomberg. But that same report notes employees haven’t seen a penny of that promised cash. And it’s not clear when it’s coming. Grok has been lagging in the AI chatbot race, with OpenAI’s ChatGPT, Microsoft’s CoPilot, Anthropic’s Claude, and Google’s Gemini leading in widespread adoption in the U.S.

The 5-minute AI newsletter. The stories that matter, in your inbox every weekday morning. Subscribe Now