MACHINE LEARNING

MERT: Self-Supervised Acoustic Music Understanding Model

By

–

04 June 2023 16h27

6/ MERT – an acoustic music understanding model with large-scale self-supervised training; it incorporates a superior combination of teacher models to outperform conventional speech and audio approaches.

Bytes Are All You Need: Direct File Classification with Transformers

By

–

04 June 2023 16h27

7/ Bytes Are All You Need – investigates performing classification directly on file bytes, without needing to decode files at inference time; achieves ImageNet Top-1 accuracy of 77.33% using a transformer backbone;

Direct Preference Optimization: Training LLMs Without RLHF

By

–

04 June 2023 16h27

8/ Direct Preference Optimization – while helpful to train safe & useful LLMs, RLHF can be complex and often unstable; this work proposes an approach to finetune LMs by solving a classification problem on the human preferences data, with no RL required.https://t.co/DZ0GSarfuT
— DAIR.AI (@dair_ai) 4 juin 2023

8/ Direct Preference Optimization – while helpful to train safe & useful LLMs, RLHF can be complex and often unstable; this work proposes an approach to finetune LMs by solving a classification problem on the human preferences data, with no RL required.

Thought Cloning: Learning to Think While Acting

By

–

04 June 2023 16h27

4/ Thought Cloning – introduces an imitation learning framework to learn to think while acting; the idea is not only to clone the behaviors of human demonstrators but also the thoughts humans have when performing behaviors.

Memory-Efficient Fine-Tuning Large Language Models

By

–

04 June 2023 16h27

5/ Fine-Tuning Language Models with Just Forward Passes – proposes a memory-efficient zeroth-order optimizer and a corresponding SGD algorithm to finetune large LMs with the same memory footprint as inference.

→ View original post on X — @mustafasuleyman

Turing Test Results: Bots Fooling Humans 40% of the Time

By

@mustafasuleyman

–

04 June 2023 13h41

turing test results: in 2m conversations, people speaking to a bot thought it was a human 40% of the time. impressive, but a long way to go https://
ai21.com/blog/human-or-
not-results
…

→ View original post on X — @miketamir

Google and OpenAI lack competitive advantages in AI

By

@miketamir

–

04 June 2023 3h01

Google "We Have No Moat, And Neither Does OpenAI" https://
bit.ly/42Jwbr7.
#AI #MachineLearning #DeepLearning #LLMs #DataScience