Building your first chatbot in Python

By: Dr. Rachael Tatman

About the speaker : In Dr. Rachael Tatman is a senior developer advocate at Rasa, where she helps developers use open source software to build great conversational AI projects. Before that she was at Kaggle (part of Google) and doing a PhD in Linguistics. If you hear a dog in the background, his name is Benson.

About the talk: This tutorial will cover the basics of Rasa, an open source library for building chatbots, including how words are translated into machine learning features and how the next conversation turn is picked. Then we'll quickly build a simple bot together and answer audience questions. .

Summary ,  Recorded
Transformers From The Ground Up

By: Sebastian Raschka

About the speaker : In recent years deep learning-based transformer models have revolutionized natural language processing. First of all, this talk will explain how transformers work. Then, we will examine some popular transformers like GPT and BERT and understand how they differ. Equipped with this understanding, we will then take a look at how we can fine-tune a BERT model for sentiment classification in Python.

About the talk: Sebastian is an Assistant Professor of Statistics at the University of Wisconsin-Madison focusing on deep learning and machine learning research. Among others, Sebastian is also a contributor to open source software and author of the bestselling book Python Machine Learning. You can find out more about Sebastian's research, books, and courses at .

Summary ,  Recorded
Find x: the missing content in data science courses

By: Abdulrahman Althobaiti

About the speaker : Abdulrahman is a data scientist at Mozn interested in solving behavioral and quantitative data problems. Passionate about building interpretable solutions, statistical inference, probabilistic models, and topology.

About the talk: Data Science courses have a tendency to provide students with multiple techniques adapted from different fields such as statistics, probability, and geometry. In this talk, we attempt to make the links between these concepts and learn how to apply them in solving business problems, confidently. .

Summary ,  Recorded
A Multi-Entity Transformer For Multi-Agent Spatiotemporal Modeling

By: Michael A. Alcorn

About the speaker : Michael is currently a Ph.D. student in Anh Nguyen's deep learning lab at Auburn University in the United States. Prior to starting his Ph.D., Michael worked for several years in industry as a Data Scientist and Machine Learning Engineer at Red Hat, and he was a prizewinner in the Research Papers Competition at the MIT Sloan Sports Analytics Conference. While at Auburn, he's had the opportunity to intern with both Adobe Research and the Cleveland Indians where he built deep learning models to address a range of problems.

About the talk: Multi-agent spatiotemporal modeling, e.g., forecasting the trajectories of basketball players during a game, is a challenging task from both an algorithmic design and computational complexity perspective. Recent work has explored the efficacy of traditional deep sequential models in this domain, but these architectures are slow and cumbersome to train, particularly as model size increases. Further, prior attempts to model interactions between agents across time have limitations, such as imposing an order on the agents, or making assumptions about their relationships. In this presentation, I will introduce baller2vec, a multi-entity generalization of the standard Transformer that can, with minimal assumptions, simultaneously and efficiently integrate information across entities and time. To test the effectiveness of baller2vec for multi-agent spatiotemporal modeling, we trained it to perform two different basketball-related tasks: (1) simultaneously forecasting the trajectories of all players on the court and (2) forecasting the trajectory of the ball. Not only does baller2vec learn to perform these tasks well, it also appears to "understand" the game of basketball, encoding idiosyncratic qualities of players in its embeddings, and performing basketball-relevant functions with its attention heads. In addition to discussing some baller2vec results, I will review two fundamental deep learning concepts behind the Transformer architecture: "embeddings" and "attention".

Summary ,  Recorded
Bayesian modeling without the math: An introduction to PyMC3

By: Dr. Thomas Wiecki

About the speaker : Thomas Wiecki is the Chief Executive Officer at PyMC Labs ( Prior to that Thomas was the VP of Data Science at Quantopian, where he used probabilistic programming and machine learning to help build the world’s first crowdsourced hedge fund. He is an author of the popular PyMC3 package — a probabilistic programming framework written in Python. He holds a PhD from Brown University.

About the talk: An Intuitive Introduction to Bayesian Modeling Bayesian modeling is an extremely powerful tool in solving data science problems across different domains. And while user-friendly modeling packages like PyMC3 exist, understanding the underlying concepts still provides a challenge for many newcomers. The main reason is that usually, statistics is taught by statisticians who provide formulas with little regard for intuition. In this talk I will take the opposite approach: throw all math out the window and explain the underlying concepts in an intuitive way.

Summary ,  Recorded


Introduction to Arabic NLP

By: Sakhar AlKhereyf

About the speaker: Sakhar Alkhereyf is a final year Computer Science doctoral student in the Natural Language Processing Lab at Columbia University, New York. His research interests include text classification incorporating social networks and Arabic NLP. Prior to his doctoral studies, he was a Research Assistant at King Abdulaziz City for Science and Technology (KACST), where he worked on different NLP and high-performance computing (HPC) projects.

About the talk: This talk will provide a brief overview of NLP, and mainly focus on NLP for the Arabic language. The Arabic language has its own challenges due to its nature as it exhibits a phenomenon called diglossia, where speakers use more than one variety of a language. In particular, people use different versions of Arabic in different situations: Modern Standard Arabic (Fusha) and Arabic spoken dialects. Also, will discuss some computational tools and datasets used for NLP in general and Arabic NLP.

Summary ,  Recorded
AI, psychology, and neuroscience: the love triangle

By:Fahad Alhazmi

About the speaker: Fahd is a Ph.D. student at the Graduate Center and Brooklyn College, City University of New York. He studies how humans learn and make decisions in complex environments using behavioral and neuroimaging methods. His work involves designing experiments, recruiting participants, and developing computational models for reinforcement learning in humans. He received his Bachelor of Science in Software Engineering from King Fahd University of Petroleum & Minerals and his Master's degree in Applied Cognition and Neuroscience from the University of Texas at Dallas. In addition to his academic interests, Fahad is also interested in data science and writing and looks forward to using both his training and experience to solve real-world challenges.

About the talk: In this talk, I aim to explore the history of the symbiotic relationships between psychology, neuroscience, and AI. While exploring this history, I will also discuss concrete examples and case studies of how insights from neuroscience are translated into the development of AI models and how recent AI advances have influenced neuroscience and psychological research.

Summary ,  Recorded
Bayesian hierarchical time series with Prophet and PyMC3

By:Matthijs Brouns

About the speaker: Matthijs is a data science lead at Xccelerated, he received a BS degree in technics, policy and management, also a MS degree in system engineering, policy analysis and management from delft university of technology.

About the talk: When doing time-series modelling, you often end up in a situation where you want to make long-term predictions for multiple, related, time-series. In this talk, we’ll see how we can combine the ideas behind bayesian hierarchical models and Facebook's Prophet package to do exactly that.

Summary ,  Recorded
Julia - a fresh approach to technical computing

By: Dr. Viral B. Shah

About the speaker: Dr. Viral B. Shah is one of the creators of the Julia language and co-founder and CEO of Julia Computing. Julia combines the ease of use of Python with the speed of C. It has been downloaded over 10 million times, and is now taught at MIT, Stanford, and many universities worldwide. The Julia co-creators were recently awarded the prestigious James H. Wilkinson prize for Numerical software.

About the talk: Julia is a modern open source programming language, originally created in 2009, and now used by over half a million users worldwide. It was designed to solve the two language problem - and provides the ease of use of Python and R with the performance of C. Today, Julia has over 4,000 community created packages, and the ability to use existing libraries written in various other languages such as Python, R, C, C++, Fortran, Java, etc. In this talk,we will be introduce the Julia language, the journey, discuss the state of the data science packages in Julia, the community’s approach to differentiable programming, and how Julia is used by enterprises worldwide.

Summary ,  Recorded
Web Scraping with Python using Requests & Beautiful Soup.

By: Juan Sebastian Vasquez

About the speaker: Juan is a Business Analyst with a digital media company focused on financial services and a Data Analytics instructor at General Assembly. He received a BS degree in Advertising with a specialization in Business from the University of Florida. Complementing his experience as a former advertiser and tech salesperson, Juan has managed to turn data into a data analyst working towards data science. He is an entrepreneur and leader who has experience with SQL, Python, business intelligence tools, GIS, digital strategy, and product development.

About the talk: we will explore web scraping basics using requests and Beautiful Soup with Python. We will be using a Google Colab Notebook and will explore the web site's HTML before jumping into any code. Aside from it being fun, web scraping can help you and your organization enrich existing data, grow new market research, dig for previously untapped insights, and unlock new narratives.

Summary ,  Recorded
Custom Scikit-learn Estimators

By: Adrin Jalali

About the speaker: Adrin is an Open Source Developer at Anaconda and a Core Developer for the Python machine learning library Scikit-learn. He has been contributing to the scikit-learn project since 2018. Before that, He was a Ph.D. student in Computational Biology at Max Planck Institute for Informatics. He worked on designing machine learning tools to classify samples collected from various biological processes.

About the talk: Often the hardest part of solving a machine learning problem can be finding the right estimator for the job. In some cases, you may decide to write your customized scikit-learn estimator, that precisely tackle the ML problem at hand. In this talk, you will learn about how to design a custom scikit-learn estimator. It will walk you through the steps to develop objects that safely interact with scikit-learn Pipelines and model selection tools.

Summary ,  Recorded
Data Management & Model Development for Autonomous Vehicles
By: Yhaya Khoja

About the speaker: Yahya is an engineer, technical consultant, and now an advisor at the Saudi Ministry of Energy. He worked as a Technical Product Manager at NVIDIA and as a Data Networks Engineer at Saudi Aramco. He received his BS in Electrical Engineering from Purdue University and his joint degree in MBA and MS in Electrical Engineering from Stanford University. Due to his expertise in managing complex projects, Mr. Khoja built a deeper understanding of ML and AI, establishing a solid foundation in business management, sales, and product development. He is passionate about developing products that leverage Machine Learning and AI to solve problems.

About the talk: utonomous Vehicles (AVs) are one of the most exciting AI applications, but it is also one of the biggest engineering challenges of our time. It is challenging because the goal is developing a vehicle than can drive safely in different geographies, weather conditions, and traffic laws. In addition, it has to decide on how to navigate when it encounters new or random situations like construction work, or a wild animal crossing the road. For this reason, developing AVs requires large amounts of data to develop numerous AI models to address different scenarios. In this talk, we will learn about the challenges, and approaches of data management and model development for AVs

Summary ,  Recorded