MACHINE LEARNING

→ View original post on X — @whats_ai

Everyone Can Master Mathematics with AI Education

By

@whats_ai

–

08 December 2023 18h01

Everyone can be great with maths.

From my interview with the fantastic Al online educator and youtuber @SerranoAcademy from @cohere.

Listen to the episode: https://t.co/sQh3ZrB0Cc #ai #math #datascience #ml #maths #whatsai pic.twitter.com/VvjpgQKYLq
— Louis-François Bouchard 🎥🤖 (@Whats_AI) 8 décembre 2023

Everyone can be great with maths. From my interview with the fantastic Al online educator and youtuber @SerranoAcademy from @cohere
. Listen to the episode: https://
youtu.be/NoyQKKQdI0M?si
=9W8PvYN37IDeDNsP
… #ai #math #datascience #ml #maths #whatsai

→ View original post on X — @miketamir

Pandas DataFrame Operations: Change and Adjust Data Sets

By

@miketamir

–

08 December 2023 17h37

pandas DataFrame Operations in Python | Change & Adjust Data Set https://
bit.ly/3tbvQkU
#AI #MachineLearning #DeepLearning #LLMs #DataScience

→ View original post on X — @ylecun

Gradients: The Core Differentiator in AI Development

By

@ylecun

–

08 December 2023 15h01

Gradients are an integral part of my job. That's what differentiates it from Elon's.

→ View original post on X — @tunguz

Exponential difficulty: AI accuracy improvements beyond human baseline

By

@tunguz

–

08 December 2023 12h36

Once you get pretty accurate, each additional percentage point of accuracy is both exponentially harder to get, and qualitatively different from the human standpoint. Don’t underestimate what they have been able to accomplish yet.

Concept Association Bias in Vision-Language Models: Contrastive vs Autoregressive

By

–

08 December 2023 10h12

We call this Concept Association Bias (CAB). We’ve found that models trained using contrastive loss (e.g. parts of BLIP and BLIP-2) also have CAB. However, models trained with autoregressive loss (e.g. OFA and BLIP-2-FlanT5) don't exhibit this bias. 4/5

CLIP’s Strategy for Handling Unmentioned Visual Concepts in Images

By

–

08 December 2023 10h12

Thus, when an image includes two concepts (e.g. a lemon and an eggplant) while the text prompt only mentions one concept (e.g. lemon), CLIP attempts to account for the unmentioned concept (like the eggplant) by saying 'purple', a color commonly associated with eggplants.
3/5

CLIP Training: Maximizing Image-Text Embedding Similarity

By

–

08 December 2023 10h12

This is because CLIP is trained using contrastive loss, where the goal is to maximize the similarity between the embeddings of images and text.
2/5

Why CLIP Misidentifies Lemon Color: Purple Instead of Yellow

By

–

08 December 2023 10h12

When you ask CLIP the color of the lemon in the image below, CLIP responds with ‘purple' instead of 'yellow' (and vice versa). Why does this happen?
1/5

→ View original post on X — @montreal_ai

Chain of Code: LLM-Augmented Reasoning and Code Emulation

By

@montreal_ai

–

08 December 2023 5h42

Chain of Code: Reasoning with a Language Model-Augmented Code Emulator Li et al.: https://
arxiv.org/abs/2312.04474 #ArtificialIntelligence #ChatGPT #LLM