(finally got around to reading in full). Amusing to read so many negative result attempts back to back to incorporate previous papers/ideas (at least in the cramming setting). Like the inline experimental result style. Like the nice code release. Like the "cramming" benchmark.
RESEARCH
-
Deep Learning Performance: Empirical Nature and High Experimental Throughput Necessity
By
–
A good display of how empirical and setting-dependent deep learning can still be, and what driving up performance looks like. In any setting it's not so much "here's how you can improve" but "here's the 10 things you should try". And why high experimental throughput is necessary.
-
Analyzing Complex Data Sets Beyond Human Capability
By
–
It is not the amount of data that is making the difference but our ability to analyse vast and complex data sets beyond anything we could ever do before.
-
The Impossibility of Parallelizing Boosting
By
–
The Impossibility of Parallelizing Boosting – investigates the possibility of parallelizing boosting. 11 of 11
-
ProGen: LLM Generates Protein Sequences with Predictable Function
By
–
ProGen – an LLM that can generate protein sequences with a predictable function across large protein families. 10 of 11
-
StyleGAN-T: New Model Enhances GAN Competitiveness for Text-to-Image
By
–
StyleGAN-T – a new model that aims to regain the competitiveness of GANs for fast large-scale text-to-image synthesis.
— DAIR.AI (@dair_ai) 29 janvier 2023
9 of 11https://t.co/leKD2VEcmvStyleGAN-T – a new model that aims to regain the competitiveness of GANs for fast large-scale text-to-image synthesis. 9 of 11
-
DetectGPT: Zero-Shot Machine-Generated Text Detection Method
By
–
DetectGPT – an approach for zero-shot machine-generated text detection. Uses raw log probabilities from the LLM to determine if the passage was sampled from it. 8 of 11
-
Open Problems in Applied Deep Learning Reference
By
–
Open Problems in Applied Deep Learning – a new reference to learn about interesting open problems in deep learning. 7 of 11
-
ClimaX Foundation Model Advances Weather Climate Science
By
–
ClimaX – a foundation model for weather and climate, including many capabilities for atmospheric science tasks. 6 of 11
-
Make-A-Video3D: New Text-to-4D Model for Dynamic Scene Generation
By
–
Make-A-Video3D – a new text-to-4D model for dynamic scene generation from input text.
— DAIR.AI (@dair_ai) 29 janvier 2023
5 of 11https://t.co/V56rEVVGYVMake-A-Video3D – a new text-to-4D model for dynamic scene generation from input text. 5 of 11