Data Phoenix Digest - ISSUE 7.2023
Upcoming webinars about unlocking data value with LLM and reducing NLP Inference costs, testing ML models for production, how to train your own LLM, 3D generation on ImageNet, universal guidance for diffusion models, CompressGPT, FriendlyCore, S-NeRF, QA4RE, and more.
Hey folks,
Welcome to this week's edition of Data Phoenix Digest! This newsletter keeps you up-to-date on the news in our community and summarizes the top research papers, articles, and news, to keep you track of trends in the Data & AI world!
Be active in our community and join our Slack to discuss the latest news of our community, top research papers, articles, events, jobs, and more...
Data Phoenix community news
Upcoming webinars:
- Unlocking Data Value with Large Language Models / June 15
- Reducing NLP Inference costs through model specialisation / June 22
Video records of past events:
- Webinar "Making AI Easy with YOLOv8"
- Webinar "Introduction to Graph Neural Networks"
- Webinar "Evaluating XGBoost for balanced and Imbalanced datasets"
Summary of the top papers and articles
Articles
Machine Learning/Artificial Intelligence Testing for Production
Most AI/ML models are now used in automation and adding decision-making but how do we know the ML model is reliable for decision-making? In this article, you will learn about a novel method for testing AI/ML models for production, including the testing workflow, metrics and tools used, and more. Take a look!
How to train your own Large Language Models
LLMs have made a significant impact in the field of AI, but most companies currently lack the ability to train these models themselves, relying instead on a few major tech firms as providers. Replit has made significant investments in developing the infrastructure necessary to train their own LLMs from scratch. In this blog post, they explain how they did it.
CompressGPT: Decrease Token Usage by ~70%
It is possible to increase the effective context window for GPT-* by asking the LLM to compress a prompt, and then feed it into another instance of the same model. But it leads to some critical issues. The author explains how to address them in an efficient manner.
FriendlyCore: A novel differentially private aggregation framework
Differential privacy (DP) ML algorithms protect user data by limiting the effect of each data point on an aggregated output with a mathematical guarantee. However, DP algorithms tend to be less accurate than their non-private counterparts. Find out how to solve this problem!
Distributed Hyperparameter Tuning in Vertex AI Pipeline
Vertex AI pipelines offer a handy way to implement end-to-end ML workflows with extremely low effort. This comprehensive article presents a new way to enable the distributed hyperparameter tuning in GCP Vertex AI pipeline. Learn more!
Papers & projects
3D Generation on ImageNet
In this paper, the authors develop a 3D generator with Generic Priors (3DGP): a 3D synthesis framework with more general assumptions about the training data, and show that it scales to challenging datasets, like ImageNet. It is based on three new ideas. Learn them!
3D-aware Conditional Image Synthesis
This paper describes a 3D-aware conditional generative model for controllable photorealistic image synthesis. It integrates 3D representations with conditional generative modeling, i.e., enabling controllable high-resolution 3D-aware rendering by conditioning on user inputs.
Universal Guidance for Diffusion Models
Typical diffusion models cannot be conditioned on other modalities without retraining. This work presents a universal guidance algorithm that enables diffusion models to be controlled by arbitrary guidance modalities without the need to retrain any use-specific components.
S-NeRF: Neural Radiance Fields for Street Views
In this paper, the authors propose a new street-view NeRF (S-NeRF) that considers novel view synthesis of both the large-scale background scenes and the foreground moving vehicles jointly. Learn more about their approach and the results of experiments!
Aligning Instruction Tasks Unlocks Large Language Models as Zero-Shot Relation Extractors
QA4RE is a framework that aligns RE with question answering (QA), a predominant task in instruction-tuning datasets. It enables LLMs to outperform strong zero-shot baselines by a large margin. This work illustrates a promising way of adapting LLMs to challenging tasks by aligning these tasks with more common instruction-tuning tasks like QA.
