MULTIMODAL AI

BLIP-2 and Visual ChatGPT: Two Advances in Computer Vision

By

–

15 March 2023 23h04

BLIP-2:
https://arxiv.org/abs/2301.12597
Visual ChatGPT:
https://arxiv.org/abs/2303.04671

BLIP-2: Efficient AI Research Using Frozen Models and Low Compute

By

–

15 March 2023 23h04

BLIP-2 is a great case study for people who want to do great works with less compute. Amazing how far using frozen models can get you! There are other similar studies such as Visual ChatGPT. VisualChatGPT indeed took it to another level.

→ View original post on X — @erictopol

LLM Achieves 85.4% on US Medical Licensing Exam

By

@erictopol

–

15 March 2023 21h54

Not everyday a large language model scores 85.4% on the US medical licensing exam (USMLE)
Implications https://
erictopol.substack.com/p/multimodal-a
i-for-medicine-simplified
…

Carnegie Mellon Multimodal Machine Learning Course Videos

By

–

15 March 2023 21h18

So good to see courses that are dedicated to this new and vibrant area of AI research. Lecture videos of Multimodal Machine Learning (MML) offered at Carnegie Mellon: https://
youtube.com/playlist?list=
PL-Fhd_vrvisNM7pbbevXKAbT_Xmub37fA
… Follow @Jeande_d for more learning resources and trends in AI research.

Multimodal Machine Learning: Fusing Vision, Audio, Text, and Actions

By

–

15 March 2023 21h18

Multimodal machine learning is a hot area in AI research. Unimodal learning has developed massively in the last 5 years. The challenge now is how we fuse different modalities(vision, audio, text, robot actions) into a single agent. GPT-4 & similar models are the beginning.

Multimodal Machine Learning Course Carnegie Mellon 2022

By

–

15 March 2023 21h18

Multimodal Machine Learning – Carnegie Mellon, 2022 A great series of lectures on multimodal machine learning(MML). The course covers fundamental concepts related to MML and recent state-of-the-art MML systems. Lectures: https://
youtube.com/playlist?list=
PL-Fhd_vrvisNM7pbbevXKAbT_Xmub37fA
… Webpage: https://
cmu-multicomp-lab.github.io/mmml-course/fa
ll2022/
…

→ View original post on X — @ronald_vanloon

VALL-E: Revolutionary Text-to-Speech AI Voice Mimicry Technology

By

@ronald_vanloon

–

15 March 2023 12h56

After ChatGPT and DALL-E, meet VALL-E – the text-to-speech #AI that can mimic anyone’s voice
by @lukekhurst @euronews Read more: https://
buff.ly/3WzDaQh #BigData #MachineLearning cc: @ronald_vanloon @yvesmulkers @pbalakrishnarao

→ View original post on X — @axeleraai

Metis AI Platform Workshop at Embedded World 2026

By

@axeleraai

–

15 March 2023 10h10

Don't miss our workshop at @embedded_world or online today at 2:30 pm! Join our VP of Product Management, Paul Neil, to learn about our Metis #AI Platform. Register now: https://
talque.com/app#/app/ngx/o
rg/DRl8NY32VVvIa2s0FYTx/session/detail/hQ03E5IfnwJZU86F4c3h
…
#EmbeddedWorld #AI #ComputerVision

→ View original post on X — @sumanth_077

GPT-4 Multimodal Capabilities: Image Input and Recipe Generation

By

@sumanth_077

–

15 March 2023 8h32

Now let's get into the details. GPT-4 is multimodal and it now accepts the images as inputs and generates captions, classifications, and analyses. Below is one such example of giving an input image of ingredients and asking GPT-4 to generate a list of recipes.