![[mldp-newsletter]-june-2023 — machine-learning-communities:-highlights-and-achievements](https://prodsens.live/wp-content/uploads/2023/07/12898-mldp-newsletter-june-2023-machine-learning-communities-highlights-and-achievements-550x367.jpeg)
[MLDP Newsletter] June 2023 — Machine Learning Communities: highlights and achievements
Let’s explore highlights and accomplishments of the vast Google Machine Learning communities over the month. We appreciate all the activities and commitment by the community members. Without further ado, here are the key highlights!
ML Training Campaigns summary
More than 35 communities around the world have hosted ML Campaigns distributed by the ML Developer Programs team during the first half of the year. Thank you all for your training efforts for the entire ML community!
- ML Study Jams: TFUG Bauchi, GDSC Uninter, TFUG Abidjan, MLAct, Universitas Pendidikan Indonesia, National Institute of Technology (Kosen), Kumamoto College, GDG Assiut, GDG Bassam, GDG Cloud Abidjan, GDG Antananarivo, Madan Mohan Malaviya University of Technology — Gorakhpur, Université d’Abomey-Calavi (UAC), ABES Engineering College — Ghaziabad, ABV-IIITM, Vishwakarma University — Pune, Pimpri Chinchwad College of Engineering and Research — Pune, GDG Cloud Edmonton, GDG Cocody, GDG Cloud Wilmington, University of Lay Adventist of Kigali
- ML Paper Reading Clubs: GalsenAI, TFUG Dhaka, Pseudo Lab, TFUG Durg, Universidad Nacional de Ingeniería, GDG Karaganda, Western University, TensorFlow Ibadan, GDG Raipur, University College Dublin
- ML Math Clubs: TFUG Dhaka, TFUG Hajipur, GDG Yangon, GalsenAI
Featured Activities by Focus Areas
Keras
Image Segmentation using Composable Fully-Convolutional Networks by ML GDE Suvaditya Mukherjee (India) is a Keras.io example that outlines how to implement a fully-convolutional network with a VGG-16 backend and use it for performing image segmentation.
KerasFuse by ML GDE Ayse Ayyuce Demirbas (Portugal) is a Python library that combines the power of TensorFlow and Keras with various computer vision techniques for medical image analysis tasks. It provides a collection of modules and functions to facilitate the development of deep learning models in TensorFlow & Keras for tasks such as image segmentation, classification, and more.
KerasCV for the Young and Restless (slides) by ML GDE Suvaditya Mukherjee (India) was an introduction about KerasCV. He discussed how basic computer vision components work; why Keras is an important tool; and how KerasCV builds on top of the established TFX and Keras ecosystem to go further. He shared his presentation at both the TFUG Malaysia and TFUG Kolkata events in June.
Complete Guide on Deep Learning Architectures Part 2: Autoencoders by ML GDE Merve Noyan (France) went through how autoencoders work and showed an implementation for Keras.
How to Train Stable Diffusion With Keras by ML GDE Derrick Mwiti (Kenya) explained how to train Stable Diffusion using KerasCV, including tips and tricks for fine-tuning using people’s images.
Stable Diffusion principle and application by ML GDE Song Lin (China) introduced Stable Diffusion principle and how to apply it with Keras.
Wrapping up Google I/O about TensorFlow & Keras (slides) by ML GDE Chansung Park (Korea) in I/O Extended hosted by GDG Daejeon was a talk sharing the newly introduced concepts of dTensor which is a very important feature in modern LLM era. He also shared basic usage of KerasCV, KerasNLP, and TFX.
Kaggle
Compete More Effectively on Kaggle using Weights and Biases by TFUG Hajipur was a meetup to explore techniques using Weights and Biases to improve model performance in Kaggle competitions. Usha Rengaraju (India) joined as a speaker and shared her insights on Kaggle and strategies to win Kaggle Competitions. She also shared tips and tricks and gave a demo related to W&B account setup and integration with Google Colab and Kaggle.
ML Olympiad for Students by GDSC Uninter was for students and ML practitioners aspirants who wanted to improve their ML experience, especially the students who attended the ML Study Jams. It consisted of a challenge of predicting US working visa applications. 320+ attendees registered for the opening event, 700+ views on YouTube, 66 teams competed, and the winner got a 71% F1-score.
On-device ML
Add Machine Learning to your Android App by ML GDE Pankaj Rai (India) at Tech Talks for Educators was a session on on-device ML and how to add ML capabilities to Android apps such as object detection and gesture detection. He explained capabilities of ML Kit, MediaPipe, TF Lite and how to use these tools. More than 700 people registered for his talk.
TensorFlow on Android for Computer Vision by ML GDE Santiago Carrillo (Colombia) at I/O Extended hosted by GDG Bogotá explained how to use the Image Labeler and Text Extractor on an Android app with ML Kit using TF Lite on Android with Firebase.
Showcasing Google MediaPipe for Audio/Video and Machine Learning by GDG Douglas shared demoing MediaPipe and how it can be run cross-platform.
LLM
#Bard ChatBard : An Intelligent Customer Service Center App by ML GDE Ruqiya Bin Safi (Saudi Arabia) is an intelligent customer service center app powered by generative AI and large language models (LLMs) using PaLM2 APIs. This demo was designed to showcase how ChatBard can revolutionize customer support.
#Bard 7 cool things you can do with Generative AI by ML GDE Tzer-jen Wei (Taiwan) was a session introducing applications and services of generative AI this time such as Generative AI Studio, Bard, ChatGPT, Imagen, Stable Diffusion, VertexAI and, etc.
https://medium.com/media/de77d16983ba0b495dcd5812db6f4d53/href
#Bard Google’s Bard Can Write Code by ML GDE Bhavesh Bhatt (India) shows the coding capabilities of Bard, how to create a 2048 game using it, and how to add some basic features to the game, such as a scoring system and a leaderboard. He is also uploading videos about LangChain in a playlist.
#Bard What is the difference between BARD and ChatGPT by ML GDE Mohamed Buallay (Bahrain) explains Bard and ChatGPT in a quick and simple video.
#PaLM #MakerSuite Generative AI with Google PaLM and MakerSuite by ML GDE Kuan Hoong (Malaysia) at Google I/O Extended George Town 2023 was a talk about a large language model with Google PaLM and MakerSuite. The event hosted by GDG George Town also included ML topics such as LLM, responsible AI, and MLOps.
#PaLM #MakerSuite Sentiment analysis with generative AI: a data-driven PaLM 2 prompt evaluation by Jason Brancazio (United States) is a simple overview of how to evaluate the quality of a sentiment analysis LLM prompt using a dataset of movie reviews, a prompt from Vertex AI’s Prompt Gallery, and the PaLM API.
#PaLM #MakerSuite Intro to Gen AI with PaLM API and MakerSuite by TFUG São Paulo was a meetup for people who want to learn generative AI and how Google tools can help with adoption and value creation. They covered a quick and easy way to start prototyping Gen AI ideas with MakerSuit and how to access advanced features of PaLM2 and PaLM API.
#PaLM Build generative AI applications with Google’s PaLM 2 model & Discover ML resources to build with from Google AI were online Zoom sessions hosted by TFUG Bassam. They explained the new AI tools from Google and how to use them.
#PaLM Hands-on with Generative AI: Google I/O Extended [Virtual] by ML GDE Henry Ruiz (United States) and Web GDE Rabimba Karanjai (United States) was a workshop on generative AI showing hands-on demons of how to get started using tools such as using PaLM API, Hugging Face Transformers, LangChain framework.
LLM Chat by ML GDE Chansung Park (Korea) is an open source software that lets you explore with different instruction-following fine-tuned LLM models in a chat interface in 2 ways (standalone Gradio and Discord bot applications).
ML Engineering (MLOps)
#Keras TFX + W&B Integration by ML GDE Chansung Park (Korea) shows how KerasTuner can be used with W&B’s experiment tracking feature within the TFX Tuner component. He also developed a custom TFX component to push a full-trained model to W&B Artifact store and publish a working application on Hugging Face Space with the current version of the model.
#KerasNLP GPT2 instruct fine-tuning pipeline by ML GDE Chansung Park (Korea) is an open source project to build a TFX pipeline for instruct fine-tuning KerasNLP’s GPT2CausalLM on Alpaca dataset. It includes some tutorial materials to show how to prepare datasets in TFRecord, how to export trained model as SavedModel, how to fine-tune GPT2CausalLM on Alpaca dataset, and how to build TFX pipeline as a whole.
#PaLM The role of ML Engineering in the time of GPT-4 & PaLM 2 by ML GDE Hannes Hapke (United States) explains the role of ML experts in finding the right balance and alignment among stakeholders to optimally navigate the opportunities and challenges posed by this emerging technology. And he gave a presentation under the same title at North America Connect 2023 and a GDG Portland event.
Responsible AI
Platform and Model Design for Responsible AI co-authored by ML GDE Sharmistha Chatterjee (India) is a new book explaining how to craft ethical AI projects with privacy, fairness, and risk assessment features for scalable and distributed systems.
Artificial Intelligence is “new electricity”: applications, challenges, ethics (slides) by ML GDE Lesly Zerna (Bolivia) talked about how to get started with AI, which projects you can follow, how to be responsible, and how to get into using Google tools. She also covered some questions from TensorFlow responsible AI explaining the pipeline for building an ML product.
ML Research
Reviews of DeepMind’s two papers by ML GDE Grigory Sapunov (UK) are shared on his social channel: 1) Model evaluation for extreme risks and 2) Faster sorting algorithms discovered using deep reinforcement learning. They covered AI safety and sorting algorithms respectively.
Guest talk By Google Engineering Fellow Blaise Aguera y Arcas hosted by TFUG Singapore. Blaise Agüera (VP and Fellow at Google Research) gave a presentation about the Cerebra project and the research going on at Google DeepMind including the current and future developments in generative AI and emerging trends.
Featured Activities by ML Frameworks
#TensorFlow In Developers Share How They Built Their Careers: From Machine Learning to Cloud, ML GDE Suvaditya Mukherjee (India) shared his experience he had earned as a GDSC lead. Now as an ML GDE, he introduced Keras and TensorFlow as his favorite tools. He commented that he loves Keras because it has an awesome open community.
#TensorFlow In How Google Enables Experts To Innovate Developer Tools From Food To Music, ML GDE Radostin Cholakov (Bulgaria) introduced TensorFlow as his favorite ML tool and also shared his NLP project, AzBuki.ML which he built using Google tools like TensorFlow and GCP.
#TensorFlow Few-shot learning: Creating a real-time object detection using TensorFlow and Python by ML GDE Hugo Zanini shows how to take pictures of an object using your webcam, label the images, and train a few-shot learning model to run in real-time.
#TensorFlow How to implement Stable Diffusion in enterprises and use TensorFlow extensions by ML GDE Hongbo Huang (China) explains how to use Stable Diffusion in enterprises, and how to modify source code through TensorFlow technologies to enable rapid deployment inference and multiplayer applications.
#TensorFlow How to Create Word Embeddings With TensorFlow by ML GDE Derrick Mwiti (Kenya) explains what word embedding is, how to represent words as numbers, and how to create word embeddings in TensorFlow.
#TensorFlow How to use TensorFlow for food recognition by ML GDE Jerry Wu (Taiwan) in Tunghai University shared about topics of OpenCV and TensorFlow in food recognition.
#TensorFlow Decision Trees: from theory to deployment on Vertex AI by ML GDE Paolo Galeone in Google I/O Extended AI explained how to train Decision Trees and Forests using the TensorFlow Decision Forests (TF-DF) package.
#Cloud In Champion Innovator Elyes Manai, based in Quebec City, Quebec, Canada, ML GDE Elyes Manai (Canada) was introduced as one of the Champion Innovators, who are technical experts in Google Cloud products and services.
https://medium.com/media/8ab759d7edfabd5266d365516b913d52/href
#Cloud #KerasCV Customizing a model using textual inversion with Vertex AI by ML GDE Pedro Gengo (Brazil) and ML GDE Vinicius Caridá (Brazil) was a session in Unraveling Google Cloud. They shared knowledge about generative AI on Google Cloud and explained how to use Keras CV in a notebook inside Workbench to add a new concept in Stable Diffusion using textual inversion.
#Cloud #Go AutoML pipeline for tabular data on VertexAI in Go by ML GDE Paolo Galeone (Italy) delved into the development and deployment of tabular models using VertexAI and AutoML with Go, showcasing the actual Go code and sharing insights gained through trial & error and extensive Google research to overcome documentation limitations.
#Cloud Beyond images: searching information in videos using AI by ML GDE Pedro Gengo (Brazil) and ML GDE Vinicius Caridá (Brazil) shared how to create a search engine where you can search information in videos. They presented an architecture where they transcript the audio and caption the frames, convert this text into embeddings and save in a vector DB to be able to search given a user query. They also introduced an architecture of how to do it using GCP.
#Cloud Generative AI Roadmap 2023 | Google Free Generative AI Certification Course by ML GDE Bhavesh Bhatt (India) is a Youtube video explaining about Google Cloud’s new courses on Generative AI.
https://medium.com/media/1a997a70e5fe811906e69af04339eeb1/href
#Cloud In Starting with new generative AI training, fast track your Google Cloud career with Innovators Plus, ML GDE Guan Wang (Singapore) shared his experience on being part of the GCP’s Innovator Champion program.
#Cloud In How to build a conversational AI Augmented Reality Experience with Sachin Kumar, ML GDE Sachin Kumar (Qatar) talked about how to build an AR app combining multiple technologies like Google Cloud AI, Unity, etc. The session walked through the step-by-step process of building the app from scratch.
#Cloud Data to AI on Google cloud: Auto ML, Gen AI, and more by TFUG Prayagraj was an event to educate students about leveraging Google Cloud’s advanced AI technologies, including AutoML and Gen AI. Googler Abirami Sukumran participated in the event as a speaker.
[MLDP Newsletter] June 2023 — Machine Learning Communities: highlights and achievements was originally published in Google Developer Experts on Medium, where people are continuing the conversation by highlighting and responding to this story.
