Generative AI - Short & Sweet 10 - MusicLM ??, DetectGPT ???, GenAI Pod?

Newsletter Sign Up?|?GAI Course?|?Sponsor

“Generative AI is a crucial step towards artificial general intelligence (AGI) in that it represents a step towards systems that are capable of generating novel and diverse outputs.”

- ChatGPT, summarising what AI researchers say.

???The Generative AI topics of the week are ...

Google focuses on GenAI, and its first Hit in 2023 is MusicLM???

MusicLM is Google Research’s awe-inspiring approach to generating music based on text only. It takes all kinds of inputs like

• rich descriptions, eg.?“Slow tempo, bass-and-drums-led reggae song. Sustained electric guitar. High-pitched bongos with ringing tones. Vocals are relaxed with a laid-back feel,and very expressive.”,?
• short prompts, eg.?“melodic techno“,?
• places, eg.?“beach in the Caribbeans “,
• hummings, instruments, skill levels of musicians, time-boxing, and much more.

The quality of the results is best in class, but not perfect yet. Get inspired here [1], see their paper [2], and take a look at the well-curated, public dataset MusicCaps [3] that MusicLM was trained on.

Detect AI-written content with DetectGPT by Stanford????

Large Language Models like ChatGPT are becoming more fluent – how can we detect if something was written by a language model or a human? Stanford University developed DetectGPT: a method for detecting if a passage was written by a particular language model.

Why does this matter? Large language models are already being used to: A) write news articles (sometimes with major errors!), B) cheat on homework [4], and more.

How does it work? DetectGPT measures the probability that a model assigns to the written text and compares that to the prob it assigns to a modification of the text. If the probability of the original is much higher than the modified text, it’s likely generated by the model.

See paper [5]. The demo is announced to come soon.?

GenerativeAI.net on the Dataframed Podcast?

Martin Musiol, the founder of generativeAI.net, took a?deep dive?into the world of generative AI. He dished on the current state, privacy, and IP worries, the best use cases, and what's to come. It's a whirlwind tour of the topic, giving us a glimpse into the future of AI and humanity. Clench your seat and enjoy the ride! [6]

Subscribing to, giving feedback about, and sharing the?newsletter?as well as our?renowned online course?will be highly appreciated and helps a lot. ??

Do you know about GenAI Gems ?? that we should consider? Or other matters??

Please, respond to this email.

Thank you for reading,?

Martin?

?

References:

[1]?https://google-research.github.io/seanet/musiclm/examples/?

[2]?https://arxiv.org/abs/2301.11325?

[3]?https://www.kaggle.com/datasets/googleai/musiccaps?

[4]?https://www.cnet.com/tech/cnet-is-testing-an-ai-engine-heres-what-weve-learned-mistakes-and-all/?

[5]?https://ericmitchell.ai/detectgpt/#?

[6]?https://www.youtube.com/watch?v=nqVpE5i80rA