YONGPENG FU

Doris apart from being an SME is also an excellent communicator.

7

This is how I founded and led a machine learning club at UBC in Year 1… 😂 It all started in Sep 2022. For 14 days, I pitched my ideas to over 35 The University of British Columbia (UBC) students across 3 faculties, convinced the VPs at our student union (AMS), and built a demo for my private version of DALL-E just for the club application. 𝐈𝐭 𝐰𝐨𝐫𝐤𝐞𝐝. UBC Applied Machine Learning and Art officially became a subsidiary of Alma Mater Society in Nov 2022. I tried running the club with my friends who knew nothing about AI. We built a website, created logos, secured $500 in funding, and even got our own office in the AMS Nest at UBC. We needed more people to do AI projects; I inspired 8 students from the UBC Computer Science Student Society with one mission: to change how artists view generative AI. I trained the team of 8 students to ethically scrape CC0 image data from 10+ domains, using tools like HTTPs, computer vision models, SSH with cloud technologies, and Python. It started off strong, but I started losing traction 3 months into the project. For the next six months, every idea I pitched, every meeting I called, every coffee invite I sent - nothing. 𝐍𝐨𝐛𝐨𝐝𝐲 𝐬𝐡𝐨𝐰𝐞𝐝 𝐮𝐩. I wanted to give up. I felt I wasn't a good fit for a President. Finding strength from my mom, I doubled down. I did something unthinkable for a Year 2 student - I convinced a Computer Science professor to join my club as our VP Project Lead. People started to attend workshops again. We 𝐠𝐫𝐞𝐰 𝐭𝐨 130+ 𝐦𝐞𝐦𝐛𝐞𝐫 𝐬𝐢𝐠𝐧𝐮𝐩𝐬, 𝐫𝐚𝐧 3 𝐰𝐨𝐫𝐤𝐬𝐡𝐨𝐩𝐬 𝐰𝐢𝐭𝐡 20+ 𝐚𝐭𝐭𝐞𝐧𝐝𝐞𝐞𝐬, and 𝐫𝐞𝐚𝐜𝐡𝐞𝐝 𝐧𝐞𝐚𝐫𝐥𝐲 20 𝐟𝐚𝐜𝐮𝐥𝐭𝐢𝐞𝐬. Soon, students with zero background in AI were explaining complex concepts with confidence. - If you've ever been told "no" to something you deeply care about, you know how it feels to wonder if they're right. But I've learned that 𝐭𝐡𝐞 𝐭𝐫𝐮𝐞 𝐭𝐞𝐬𝐭 𝐨𝐟 𝐲𝐨𝐮𝐫 𝐯𝐢𝐬𝐢𝐨𝐧 𝐢𝐬 𝐧𝐨𝐭 𝐡𝐨𝐰 𝐨𝐭𝐡𝐞𝐫𝐬 𝐬𝐞𝐞 𝐢𝐭 - 𝐢𝐭'𝐬 𝐡𝐨𝐰 𝐲𝐨𝐮 𝐟𝐢𝐠𝐡𝐭 𝐟𝐨𝐫 𝐢𝐭 𝐰𝐡𝐞𝐧 𝐧𝐨 𝐨𝐧𝐞 𝐞𝐥𝐬𝐞 𝐰𝐢𝐥𝐥. #UBC #Leadership #Founder #Resilience #ComputerScience

194

7 Comments

👉🏼 ChatENT: Augmented Large Language Model for Expert Knowledge Retrieval in Otolaryngology-Head and Neck Surgery 🤓 Cai Long 👇🏻 https://lnkd.in/e7_3F-hi 🔍 Focus on data insights: - ChatENT outperformed ChatGPT4.0 in both the Canadian Royal College OHNS sample examination questions challenge and the US board practice questions challenge. - A significant error reduction of 58.4% and 26.0% was observed, respectively. - ChatENT demonstrated enhanced performance in analyzing and interpreting OHNS information, generating fewer hallucinations and showing greater consistency. 💡 Main outcomes and implications: - ChatENT is the first specialty-specific knowledge retrieval artificial intelligence in the medical field utilizing the latest LLM. - Promising applications in medical education, patient education, and clinical decision support are highlighted. - Overcoming limitations of existing LLMs, ChatENT signals a future of more precise, safe, and user-friendly AI applications in OHNS and other medical fields. 📚 Field significance: - Artificial Intelligence - Medical Education - Clinical Decision Support 🗄️: [#Otolaryngology #ArtificialIntelligence #MedicalEducation #ClinicalDecisionSupport]

1

I'm excited to share that I've completed the "Machine Learning in Production" course! It provided a clearer and more organized understanding of how to effectively deploy and scale ML models in real-world scenarios. A big thanks to Canadian Tire Financial Services for supporting my personal development. #MachineLearning #ProfessionalDevelopment #AI

39

3 Comments

Unlock the full potential of your Large Language Models (LLMs) with Intel® Extension for PyTorch (IPEX) and the Intel® LLM Library for PyTorch (IPEX-LLM). Download this whitepaper to explore the optimization of LLMs, including their performance, resource utilization, and response times in real-world applications. Link - https://intel.ly/3BBJ4ey

1

📚 Quick review of 'Ace The Data Science Interview' by Kevin Huo and Nick Singh 📕🐒 I've found this book at just the right moment in my career journey! As a professional seeking new opportunities in Canada, I've often felt lost and questioned if I was on the right track. This book has been incredibly helpful so far, even before diving into the technical interview section. The resume chapter, in particular, has been a game-changer. After revising my resume 20+ times, I still felt something was off. However, this book, along with posts about polishing resumes by Chong Lee, has finally helped me get on the right track. A huge thank you to Kevin Huo, Nick Singh, and Chong Lee for your invaluable guidance and support! While the book is focused on the US job market, I've found its insights useful for my Canadian job search. I've enjoyed reading this book and am already looking forward to the next chapter! 🤓📖 #DataScience #JobSearch #BookReview

7

1 Comment

I found this article from the University of British Columbia (UBC) that I wanted to share with my network! It discusses UBC's exploration of generative AI (GenAI) tools in teaching and learning. The article highlights the potential of GenAI to enrich the educational experience, from sparking creativity to supporting personalized study habits. However, UBC emphasizes the importance of responsible use, addressing issues like plagiarism and potential biases within AI models. This is a fascinating look at how UBC is embracing innovation while ensuring academic integrity. Additionally, UBC has included a helpful flowchart adapted from the "ChatGPT and Artificial Intelligence in Higher Education: Quick Start Guide" (UNESCO and UNESCO International Institute for Higher Education in Latin America and the Caribbean, 2023). This flowchart can be a resource for students considering using generative AI tools for academic purposes. It's important to remember that the flowchart is a guide and doesn't replace specific course or program policies on GenAI use. The flowchart itself was created in June 2023 and is currently under review by UBC. Link: https://lnkd.in/eT2QGMaP #GenerativeAI #Education #UBC #Innovation

23

1 Comment

I recently attended the AI Engineering bootcamp via AI Makerspace as a part of Toronto Machine Learning Society (TMLS), where I delved into advanced topics like Retrieval Augmented Generation (RAG), the Reasoning-Action (ReAct) pattern, and agentic behavior in AI. A key highlight was building an Open-Source Agentic RAG Application using @Llama-8B-instruct, LangChain, and LangSmith, featuring LangGraph for stateful and autonomous agent cycles. We utilized tools such as the Snowflake-arctic-embed-m embedding model, Qdrant vector database, ragas and OpenAI GPT-4o for evaluation and synthetic test data generation. This hands-on experience enhanced my skills in Prompt Engineering, Fine-Tuning, and using Function Calling for tool access. I'm excited to apply these cutting-edge techniques to real-world challenges! #AI Makerspace #LLM #LangChain #TMLS #AgenticAI #RAG #Llama3

57

3 Comments

Are you curious about #AI and #ML insights on #thunderstorms? 🌩️   If so, let's explore our latest #paper: “Machine Learning Investigation of Downburst-Prone Environments in #Canada” Access the article: https://lnkd.in/gpkbDwhP Journal of Applied Meteorology and Climatology, vol. 63, p. 677-697, 2024. • Binary classified thunderstorms with/without #downbursts. • Identified superior performance of the #random_forest model. • Outlined #resampling methods' pros/cons for dataset balance. • Highlighted ML model's learning of input #convective trends. • Found higher importance for #thermodynamic parameters. Gratitude to Dr. Djordje Romanic, for his priceless supervision and support during this #research. Don't miss out on diving into the fascinating intersection of #data and #wind #science! You'll be amazed!

33

3 Comments

To Label or to Pseudo Label? Active Learning vs Semi-Supervised Learning: I'm thrilled to share that as H3M Analytics, our Mitacs collaboration with the University of Ottawa has resulted in our first paper being published in this project at the The 37th Canadian Conference on Artificial Intelligence! The paper, co-authored by Bahar Afshar, Paula Branco, Utku Gorkem Ketenci, and Hikmet Mazmanoglu, explores the effectiveness of Active Learning and Semi-Supervised Learning methods for Windows malware detection. Using the Microsoft malware prediction dataset, we found that both methodologies enhance detection accuracy, especially in scenarios with limited labeled data, with Semi-Supervised Learning showing superior outcomes in certain settings. For more details, check out the paper here: https://lnkd.in/eK2gTE7d #Research #Innovation #AI #Compliance

19

2 Comments

If you didn't have a chance to attend my Getting Started with Generative Text & Fine-tuning LLMs in Hugging Face workshop at Toronto Machine Learning Society (TMLS) Toronto Summit last month, all of the content is now available online! 🚀 Highlights of the workshop include: 💡 An intro to #largelanguagemodels and generative text models in #HuggingFace 🤗 🌡️An explanation of different decoding strategies for #LLMs (temperature, top-p, top-k, beam & greedy search) ⚙️Fine-tuning GPT-2 medium to talk like a Jedi Master 😱Using LoRA to fine-tune GPT-XL using only a Google Colab free instance 🙊Throwing shade at AGI, in-context learning, and George Lucas Check out the full 90 minute video of the workshop here 📽️: https://lnkd.in/gh5aXy4Q View the workshop materials on Github🧑‍💻: https://lnkd.in/gb7aKSCU Give the code a spin for yourself in Google Colab 🧪: https://lnkd.in/gvnpZWSu

11

🌸🍒 Calling all cherry blossom enthusiasts in Vancouver, It is the season! 🍒🌸 Curious about where to find the most trees in our beautiful city? 🌳 Look no further! 🗺️ According to my analysis using Python, the top three neighborhoods boasting the highest number of trees are Kensington-Cedar Cottage, Renfrew-Collingwood, and Hastings-Sunrise. 🏡 But when it comes to cherry trees stealing the spotlight, a different neighbourhood takes the crown! 🏆 Mount Pleasant reigns supreme with the greatest number of cherry trees🌸 Can you imagine the breathtaking sight in spring? 🌸 What's even more exciting? 🎉 We're treated to a colorful cherry blossom display from February to June, thanks to different varieties like Akebono and Kwanzan. 🌸 Akebono blooms first, followed by Kwanzan a week or two later, ensuring a prolonged period of floral delight! 🌸 For this delightful analysis, I enlisted the help of python (Altair and Pandas). Check out the full breakdown here: https://lnkd.in/gFyXGS2F 📊 https://lnkd.in/gPSbPbn2 #CherryBlossoms #VancouverBlooms #DataAnalysis #CityOfVancouver #Python

36

2 Comments

Toronto Machine Learning Society (TMLS) #tmls2024 Incredible talk from Patrick Tammer of Scale AI “Successfully Integrating AI in your Strategy and Business Operations – Lessons Learnt from Investing” Lots of learnings about the Canadian AI ecosystem, also most importantly how to define an AI strategy and how to implement it. #aistrategy #datastrategy #aicompetency #startups

12

2 Comments

🚀 Exciting News! 🚀 I'm thrilled to announce a fun open source project that my colleague Jane(Jingyan) Jiang and I have developed: the **Cold Email Writer for Research Opportunity ** on GitHub! 🎉 Check it out here: [GitHub - Research Opportunity Cold Email Writer](https://lnkd.in/e6bDPvvq) Why did we create this tool? 🤔 We wanted to help students and researchers easily reach out to professors with thoughtful and personalized emails. Based on our personal experience, we know crafting the perfect email can be daunting, especially when trying to make a great first impression. Our tool aims to simplify this process and enhance the quality of these important communications. How does it work? 🤖 We compared outputs from multiple large language models (LLMs) and decided to use the DeepSeek Large Language Model. By iteratively improving our prompt engineering, we've ensured that the final email generated is as desired as possible. This project is open for feedback and all kinds of collaborations through GitHub. 👩‍💻 👨‍💻 If you have any brilliant ideas or suggestions, please feel free to reach out to us, fork the repo and submit pull requests (details are listed in ReadMe!). We're excited to see how this project evolves with your contributions! 🌟 #OpenSource #Research #Collaboration #AI #DeepLearning #PromptEngineering

67

6 Comments

FlexAttention: Simplifying Attention in PyTorch _________________________________________________________________ FlexAttention is a new PyTorch API that combines the flexibility of PyTorch with the performance of optimized attention methods. It allows researchers to implement various attention mechanisms easily without needing to write custom kernels. Key Points: - Flexibility with Performance: -- Traditional attention methods offer high performance but limited flexibility. -- FlexAttention provides a flexible API that supports many attention variants with minimal effort. - How It Works: -- Standard Attention: Calculates attention scores using queries, keys, and values, normalizes them, and computes the weighted sum of values. -- FlexAttention: Allows for modifications to the attention scores through a flexible scoring function, enabling customization for different types of attention mechanisms. - Examples of Attention Variants: -- Relative Position Encoding: Adds positional information to attention scores based on relative positions. -- ALiBi Bias: Applies a bias to attention scores to account for position differences, enhancing model performance. -- Causal Mask: Masks future positions in the sequence to ensure the model only attends to previous positions, useful for autoregressive models. Handling Sparsity: -- BlockMask: Efficiently manages sparse attention masks, such as those used in causal attention, by creating masks that prevent unnecessary computations. Learn More: https://lnkd.in/g58pxB4A

8