RESEARCH

→ View original post on X — @terence_mills

The 12 Biggest AI Mistakes You Must Avoid

By

@terence_mills

–

25 June 2023 17h01

The 12 Biggest AI Mistakes You Must Avoid
#AI #AIio #BigData #ML #NLU #Futureofwork http://
ow.ly/xB8f30svSee

→ View original post on X — @abacusai

MIT Launches Free Introduction to Deep Learning Course

By

@abacusai

–

25 June 2023 17h00

For those who want to learn Deep Learning: "Introduction to Deep Learning" is a new FREE course just released by MIT. You can stream it directly from YouTube: https://youtube.com/playlist?list=PLtBw6njQRU-rwp5__7C0oIVt26ZgjG9NI Great opportunity!

→ View original post on X — @rasbt

Critique des problèmes méthodologiques dans l’évaluation GPT-4

By

@rasbt

–

25 June 2023 16h52

> "what hit most was that it got uploaded without coauthor permission" Sure, this was another issue. But wasn't the core issue the flawed evaluation setup the authors used? Using GPT-4 to score itself? And repeated prompting until the answer was correct, thus reaching 100%?

→ View original post on X — @rasbt

Critique of Flawed Research Paper and Inadequate Apology

By

@rasbt

–

25 June 2023 16h35

"[…] could have been a very interesting and valuable paper had the data been collected with consent" So, if people gave consent for using the data, the flawed paper would become interesting and "valuable"? Long story short: Problematic paper, problematic apology letter.

→ View original post on X — @rasbt

Dataset Consent vs Flawed Evaluation: Main Issues Debate

By

@rasbt

–

25 June 2023 16h25

Hm, it seems the letter largely highlights the issues regarding dataset consent. Of course, this is an issue. But wasn't the main issue of this study the flawed evaluation and, consequently, the wrong claims?

Wanda: Efficient LLM Pruning Without Retraining

By

–

25 June 2023 15h02

9/ Wanda – introduces a simple and effective pruning approach for LLMs; the approach requires no retraining or weight update and outperforms baselines of magnitude pruning.

AudioPaLM Multimodal LM Enables Zero-Shot Speech Translation

By

–

25 June 2023 15h02

10/ AudioPaLM – fuses text-based & speech-based LMs, PaLM-2 and AudioLM, into a multimodal architecture; outperforms existing systems for speech translation tasks and has zero-shot speech-to-text translation capabilities.

LMFlow: Extensible Toolkit for Foundation Model Fine-tuning

By

–

25 June 2023 15h02

7/ LMFlow – an extensible toolkit that simplifies finetuning and inference of general large foundation models; supports continuous pretraining, instruction tuning, parameter-efficient finetuning, alignment tuning, and large model inference.

LOMO: Memory-Efficient Optimizer for Full LLM Parameter Tuning

By

–

25 June 2023 15h02

5/ LOMO – proposes a new memory-efficient optimizer that combines gradient computation and parameter update in one step; enables tuning the full parameters of an LLM with limited resources.