Speaker: Chia-Chi Chang https://docs.google.com/document/d/1LwQG8pLLO2PEviExoU3xiqyclnaT9kfhwsCGVKVRB2I/edit?usp=sharing Although there are several data mining tools in python, you can use them to deal with almost every kind of data (numeric, text, image, audio, ...) you met. Besides, there are also lots of modeling tools in python, you can use them to build the FIRST LIGHTING MODEL to solve the problems. However, if you want to solve problems deeply, most of time, you need to write down the customerized models and solve them by yourself. Instead of using fast modeling tools, you need to know more about the essential things in modeling: What is a model ? - How models solve your problems ? - What is the connection between models and data ? - What is the important data ? important model ? I deeply believe ... The more the connection between models and data you know, the deeper the problem you can solve. Outline: What is Modeling ? Data, Model, Evaluators Direct Problem: Data + Evaluators → Model Inverse Problem: Data + Models → Evaluator Hacking Models with Metric Learning Data as a Model & Model as a Data Duality between Dimension Reduction and Clustering ... https://www.youtube.com/watch?v=PznPp-BbwyU

?美國的會眾比台灣的會眾更主動搭話 參加 Congerence 的時候，你是那個躲在角落的小孤獨嗎？ 會場人來人往，放眼望去都是陌生人，內向者到底要怎麼開始交流呢？ 議程跑跳經驗豐富的 TP 鼓勵大家主動的開啟話題，一起營造出開放友善的對話環境 雖然乍看之下我們的日常都很相似，只要再近一步了解也許就會找到有趣連結！ 今年九月，我們相約在 PyCon TW 見囉～！ - ? 本集節目感謝 PyCon TW 2023 金級贊助商 Netskope 贊助 - ?點擊收聽全集：https://reurl.cc/WG7XKx Follow “PyCon Taiwan” ⭐️ Official Website: https://tw.pycon.org ⭐️ Facebook: https://www.facebook.com/pycontw ⭐️ Instagram: https://www.instagram.com/pycontw ⭐️ Twitter: https://twitter.com/PyConTW ⭐️ LinkedIn: https://www.linkedin.com/company/pycontw ⭐️ Blogger: https://pycontw.blogspot.com ... https://www.youtube.com/watch?v=HeQ89qatviM

Day 1, R1 15:55–16:10 Productization of machine learning (ML) solutions can be challenging. Therefore, the concept of operationalization on machine learning (MLOps) has emerged in the past few years for effective model lifecycle management. One of the core aspects of MLOps is "monitoring". ML models are built by experimenting with a wide range of datasets. However, since the real data continues to change, it is necessary to monitor and to manage model usage, consumption, and results of models. MLflow is an open-source framework designed to manage the end-to-end ML lifecycle with different components. In the talk, the basic concepts of MLflow will be introduced. Then, MLflow Tracking will be the main focus. You will know how to track experiments for recording and comparing parameters and results by MLflow Tracking. Slides: https://speakerdeck.com/sucitw/track-machine-learning-applications-by-mlflow-tracking Speaker: Shuhsi Lin A data engineer and python programmer. Currently working on various data applications in a manufacturing company. Research interests: IoT applications, data streaming processing, data analysis and data visualization. ... https://www.youtube.com/watch?v=76QWG9di1Hs

PyCon APAC 2022｜一般演講 Talks｜國泰金控 Cathay Financial Holdings / 美光科技 Micron 冠名贊助 ✏️ 共筆 Note：https://hackmd.io/@pycontw/rJ5aL67yo ?? Slido：https://app.sli.do/event/ek2NTzkytzYkU1XVkpvdzk ? 語言 Language：英文 English ? 層級 Level：中階 Intermediate ? 分類 Category：應用 Application ? 摘要 Abstract ? Finding a working shift schedule can be a tedious task, there’s usually various requirements to juggle all at once, from time-offs, shift types, conflicting schedules, to personal preferences; and one may easily miss a requirement or two by mistake. Instead, we can use so called “solvers” and leave the hard lifting up to computer to find not just a valid schedule, but a good one; and as it turns out, the same technique also works for a broad spectrum of problems such as conference scheduling, vehicle routing, bin packing, and more. ? 說明 Description ? This talk will likely interest those who likes to explore using Python to solve logical problems, because in some way, finding a shift schedule is a lot like playing Sudoku. There’s a lot of holes that you need to fill, but only certain element can fit in the holes. So if we can write programs that solves Sudoku, then surely we can write programs that produces a good shift schedule, right? Indeed we can, but there’s an even better approach. Instead of writing our own custom program that specifically schedules shifts, we can use off-the-shelve solvers that are very good at solving a more general kind problem, and transform our original problem into a form that it understands. Behind the scene these off-the shelve solvers use similar technique used by the Sudoku solver mentioned above (e.g. backtracking and constrain propagation), albeit much more advanced, since tons of optimization and clever tuning have went into make them work well. By then end of the talk, you’ll know the basic of Z3’s Python API, how to formulate the shift scheduling problem in propositional logic, and a slight idea of how it can be applied to solve similar problems such as conference scheduling, vehicle routing, bin packing, and more. Audience are suggested to have a good understand of list comprehension and itertools since both will be heavily used throughout the talk. Ideally, some experience with predicate logic is preferred, though it is not strictly necessary. However, no domain knowledge is required. - Side note: Z3 is an immensely powerful tool; this talk can only covers a tip of iceberg in terms of what it can do. Other interesting use of Z3 includes: software verification, program synthesis, exploit generation, code simplification, etc. ? 講者介紹 About Speaker - Shung-Hsi Yu ? Kernel Engineer at SUSE working on BPF #pycontw #pyconapac2022 #python #shiftschedule #algorithm #z3 Follow “PyCon Taiwan” ⭐️ Official Website: https://tw.pycon.org ⭐️ Facebook: https://www.facebook.com/pycontw ⭐️ Instagram: https://www.instagram.com/pycontw ⭐️ Twitter: https://twitter.com/PyConTW ⭐️ LinkedIn: https://www.linkedin.com/company/pycontw ⭐️ Blogger: https://pycontw.blogspot.com ... https://www.youtube.com/watch?v=T6Q2fPnPgUU