[MLDP Newsletter] April 2023 — Machine Learning Communities: highlights and achievements

[MLDP Newsletter] April 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 Campaign

ML Study Jams

More than 22 communities around the world have hosted ML Study Jams and have guided beginners into the world of machine learning!

• GDSC Uninter hosted ML Study Jam Professionals #1 , #2, and #3 for students aspiring to become ML practitioners. More than 2500 participants studied ML together for 4 weeks. Check out the group’s ML Study Jam playlist. ML GDE Vinicius Caridá (Brazil) participated in the second event and gave a talk about how ML solves real problems for the bank and its customers and his journey to reach the top of his data career.
• TFUG Bauchi hosted ML Study Jams 2.0 to introduce the most basic ML concepts for beginners.

Featured Activities by Focus Areas

Keras

[0.11] keras starter: unet + tf data pipeline by ML GDE Aritra Roy Gosthipaty (India) is a new starter guide for the Vesuvius Kaggle competition. He also shared When Recurrence meets Transformers, the implementation of Temporal Latent Bottleneck Networks, proposed in the paper.

StableDiffusion — Textual-Inversion implementation app by ML GDE Dimitre Oliveira (Brazil) serves as an example of how to implement code from research and fine-tunes it using the Textual Inversion process and also shows relevant use cases for valuable tools and frameworks like HuggingFace, Gradio, TensorFlow serving, and Keras-CV.

In Understanding Gradient Descent and Building an Image Classifier in TF From Scratch (photo), ML GDE Tanmay Bakshi (Canada) talked about how to develop a solid intuition for the fundamentals backing ML tech, and actually build a real image classification system for dogs and cats, from scratch, in TF.Keras.

TensorFlow and Keras Implementation of the CVPR 2023 paper by Usha Rengaraju (India) is a research paper implementation of BiFormer: Vision Transformer with Bi-Level Routing Attention.

Smile Detection with Python, OpenCV, and Deep Learning by Yacine Rouizi is a tutorial explaining how to use deep learning to build a more robust smile detector using TensorFlow, Keras and OpenCV.

On-device ML

Adding ML to our apps with Google ML Kit and MediaPipe by ML GDE Juan Guillermo Gomez Torres (Bolivia) introduced ML Kit & MediaPipe, and the benefits of on-device ML. In Startup Academy México, he shared how to increase the value for clients with ML and MediaPipe.

Mobile Apps with TFLite by TFUG Chandigarh was an offline event to make students familiar with TFLite and its use cases. It was organized in collaboration with GDSC MMDU, Mullana, Haryana to share knowledge about mobile apps development with TFlite, making models and deployment with introduction to LLMs.

ML Engineering (MLOps)

Serving With TF and GKE: Stable Diffusion by ML GDE Chansung Park (Korea) and ML GDE Sayak Paul (India) discusses how TF Serving and Kubernetes Engine can serve a system with online deployment. Stable Diffusion is one of the examples that TF and GKE can serve with online deployment. They broke down Stable Diffusion into main components and how they influence the subsequent consideration for deployment. Then they also covered the deployment-specific bits such as TF Serving deployment and k8s cluster configuration.

ML Infra and High Level Framework in Google Cloud Platform by ML GDE Chansung Park (Korea) delivered what MLOps is and why it is hard; why cloud + high level framework (TFX) is a good starter and how TFX is seamlessly integrated with Google Cloud (Vertex AI and Dataflow). He also shared some use-cases from the past projects that he and ML GDE Sayak Paul (India) have done in the last 2 years.

Open and Collaborative MLOps by ML GDE Sayak Paul (India) talked about why openness and collaboration are two important aspects of MLOps. Then he gave an overview of Hugging Face Hub and how it integrates well with TFX to promote openness and collaboration in MLOps workflows.

April ML meetup with TFUG Kolkata by TFUG Kolkata was a monthly meetup for ML professionals. People discussed how ML can be integrated in their workflows and shared best practices in MLOps. ML GDE Ayush Thakur gave a talk at the event.

Kubernetes Good Practices / MLOps Night on VertexAI by GDG Cloud Paris shared 16 best Kubernetes practices and how to implement an MLOps platform on VertexAI.

ML Research

Learning JAX in 2023: Part 3 — A Step-by-Step Guide to Training Your First Machine Learning Model with JAX by ML GDE Aritra Roy Gosthipaty (India) and ML GDE Ritwik Raha (India) shows how JAX can train linear and nonlinear regression models and the usage of PyTrees library to train a multilayer perceptron (MLP) model.

Paper review: Power-seeking can be probable and predictive for trained agents by ML GDE Grigory Sapunov (UK) explains the details of the new article on AI Safety.

Opening Pandora’s box: understanding the paper that revolutionized the field of NLP (video) by TFUG São Paulo, ML GDE Vinicius Caridá (Brazil), and ML GDE Pedro Gengo (Brazil) was an event sharing the secret behind the famous LLM and other generative AI models. To understand the-state-of-the-art technologies, they studied the paper, Attention Is All You Need, and learned the full potential that the technology can offer.

Featured Activities by ML Frameworks

#TensorFlow Training a recommendation model with dynamic embeddings by ML GDE Thushan Ganegedara (Australia) explains how to build a movie recommender model by leveraging TensorFlow Recommenders (TFRS) and TensorFlow Recommenders Addons (TFRA). The primary focus was to show how the dynamic embeddings provided in the TFRA library can be used to dynamically grow and shrink the size of the embedding tables in the recommendation setting.

#TensorFlow How I built the most efficient deepfake detector in the world for $100 by ML GDE Mathis Hammel (France) was a talk exploring a method to detect images generated via ThisPersonDoesNotExist.com and even a way to know the exact time the photo was produced.

#TensorFlow TensorFlow Model Card Toolkit 2.0 Release by ML GDE Hannes Hapke (United States) is the first fully OSS release run by the MCT community.

#TensorFlow Tech Talks 2.0: Introduction to Machine Learning with TensorFlow by Aritra Roy Gosthipaty (India) was for teaching faculties, trainers, and mentors to bring the latest of Google technology in their learning environment. He covered why TensorFlow and how to navigate through it.

#TensorFlow #VertexAI Setup of NVIDIA Merlin and Tensorflow for Recommendation Models by ML GDE Rubens Zimbres (Brazil) presented a short review of Recommendations Algorithms as well as the Two Towers algorithm, and setup of NVIDIA Merlin on premises and on Vertex AI.

#TensorFlow Training a language model with 🤗 Transformers using TensorFlow and TPUs co-authored by ML GDE Sayak Paul (India) and his colleague showed how to train a language model from scratch using TensorFlow, Hugging Face Transformers with TPUs.

#TensorFlow Build the next generation of Web Apps with TensorFlow.js by GDG Aveiro shared how to use visual computing to recognize people, pets, and objects in a web application and how to manage application scalability and security. They also explored how to use TensorFlow to deploy and manage Web Apps.

#Vertex AI In Monitoring the Quality of AI Driven Search, ML GDE Gad Benram (Portugal) talked about the ways that Google is improving search and tools on GCP to build AI search solutions such as Embedding Endpoints, Matching Engine, Vertex AI and more. In Lean Search Architectures on GCP, he demonstrated the versatility of Google Cloud as a platform to support the development and hosting of ML-based searching solution systems.

#BigQuery Optimizing Costs in BigQuery by ML GDE Jéssica Costa (Brazil) focused on bringing some tips for optimizing costs in BigQuery, mainly regarding queries.

#VertexAI #BigQuery Hands-on ML with Google Cloud — a Gameathon by GDG Cloud Kolkata shared how to create quick and easy ML models with Vertex AI and BigQuery.

#Vertex AI What’s inside Google’s Generative AI Studio? by ML GDE Gad Benram (Portugal) shared the preview of the new features and what you can expect from it. In How to pitch Vertex AI in 2023, he delivered the six simple and honest sales pitch points for Google Cloud representatives on how to convince customers that Vertex AI is the right platform.

#Vertex AI Organize, store, and serve your ML features with Vertex AI Feature Store by GDG Cloud Milano explored the benefits of using a feature store, and how Vertex AI Feature Store can help you to build, deploy, and manage machine learning models more efficiently.

#Dialogflow Integrating CX Dialogflow into different channels by GDG Madrid focused on how to use Dialogflow CX to build chatbots and virtual assistants, and how to create integrations for different messaging channels like Telegram, WhatsApp, Facebook Messenger.

#GCP Machine Learning on Google Cloud Platform by ML GDE Nitin Tiwari (India) was a mentoring session aiming to provide students with an in-depth understanding of the processes involved in training an ML model and deploying it using GCP. In Building robust ML solutions with TensorFlow and GCP, he helped students understand how to leverage the capabilities of Google Cloud Platform and TensorFlow to build robust ML solutions and deploy custom ML models.

#GCP Getting Started with AI on Google Cloud Platform: Episode 2: Exploring Google Cloud’s Machine Learning APIs by ML GDE Youssef Ben Dhieb (Tunisia) explored ML APIs.

#Cloud Google Cloud Community Days 2023 — Mumbai by GDG Cloud Mumbai and Google Cloud Community Day Madurai 2023 by GDG Cloud Madurai were conferences hosted by communities. Those events aimed to help people understand the latest innovations and trends in the cloud computing domain. They offered a diverse range of sessions such as machine learning, artificial intelligence, infrastructure, data, and more.

#Cloud ML-A-THON, a competition conducted by GDSC TKMCE & GDG Cloud Kochi was a hackathon where each team is required to predict test datasets by training ML models using datasets provided in Kaggle.

#Bard Get to the bottom of it and ask the Bard by ML GDE Yucheng Wang (China) tested Bard with various questions and shared the results he received.

#Bard Google Bard vs ChatGPT: The AI Battle You’ve Been Waiting For! by ML GDE Bhavesh Bhatt (India) compared the capabilities, accuracy, and overall performance of these two cutting-edge models to determine which one reigns supreme.

#Bard Bard can now code and put that code in Colab for you by ML GDE Sam Witteveen (Singapore) showed how Bard makes code.

#PaLM Language models which PaLM can speak, see, move, understand by GDG Cloud Taipei was an event for those who want to understand the concept and application of PaLM. They shared the concept and application of PaLM and main characteristics, functions, and etc.

Others

In Developer Journey: April 2023, ML GDE Ruqiya Bin Safi (Saudi Arabia) was introduced as an ML expert who has been building groundbreaking ML models in image and speech recognition.

The secret sauce to creating amazing ML experiences for developers by ML GDE Gant Laborde (United States) was a podcast sharing his ‘aha moment’, 20-year experience in ML, and secret to creating enjoyable and meaningful experiences for developers.

#MLOlympiad: Mentorship Sessions by TFUG Saudi & ML GDE Ruqiya Bin Safi (Saudi Arabia) was preceded along with Dialect_Recognition competition, discussion session, and ceremony & discussion of the winner’s solution.

Interacting with LLMs in a post-chatGPT world by TFUG Singapore, Augmenting Large Language Models by ML GDE Martin Andrews (Singapore) and Autonomous AI + Advanced LangChain by ML GDE Sam Witteveen (Singapore) covered recent techniques that are showing promise in the prompt-engineering and fine-tuning area of LLMs. Sam looked at some Autonomous AI with LLMs systems and how you can build similar things in LangChain.

[MLDP Newsletter] April 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.