Menu
Log in


Log in

AI on Pi Day - 2022 Edition

The 2022 event

AIDAUG will hold its first planetary-wide event on Pi Day (πday) 2022. This AI marathon features:

  • A major keynote featuring a word-renown AI leader: Dr. Seth Dobrin, Chief AI Officer, IBM, 
  • Several sessions of 30 minutes to ensure a very dynamic event. Sessions should last 20 minutes of lecture and 10 minutes of Q&A, as we want to create a strong network.
  • Download the planning
  • If you have registered head over to this news-post to see Links for Zoom and Slack Channel here (for registered users only!)

Major dates

March 14th The event starts - see schedule below. Make sure you are a member of register to the event to get the web conference links. 
February 28th  Speakers get notification of acceptance of papers.
February 14th  Call for Papers is now closed.
January 17th  Call for Papers opens.

Registration

Registration is free for all AIDAUG members. You do not have to do anything, simply join AIDAUG at http://aidaug.org/join.

Day schedule 

 Length Title   Speaker
 (min)    
15  Introduction & Welcome  Jean-Georges Perrin
 30 What is AIDAUG?  Ralf Roeber
 30 Boost your financial planning with automatic forecasting  Anders Kramer Knudsen
 30 Perfecting the last mile of the analytics supply chain  Ryan Dolley
 30 Forecasting at Scale – an Automated Approach

 David Trisl

 30 Why and When should I use IBM Business Analytics vs. PowerBI?  Rikke Jacobsen
 30 Tethering AI with Blockchain  Karen Kilroy
 30 AI for Agave  Eduardo Ulises Moya Sánchez
 30 Why is data quality important in an NLP project?  Pierre-Nicolas Perrin
 60 Human Centered AI, keynote by Seth Dobrin  Seth Dobrin
 30 Informix data analysis by Cloud Pak for Data  Jan Musil
 30 Is it time to switch to Linux on your desk?  Michael Mikowski
 30 Detecting Icy Bridges with AWS, Image Processing and AI  John Fred Davis
 15Closing statementJean-Georges Perrin, Ralf Roeber, and Stuart Litel

Boost your financial planning with automatic forecasting

How to implement financial planning models which leverage historical data to forecast the next periods. These models include internal as well as external data sources to the forecast process. I will precent ways to model effects from past data. This includes classical forecasting techniques such as exponential smoothing and ARIMA modelling. But also, bring in Fourier transformations, to enrich data understanding.

For larger and more complex scenarios neural network models are introduced to handle very large numbers of time series, that needs to be forecasted.

The theory is coupled with two implementations of forecast models, that has been implemented using IBM TM1 planning software together with open source tools to predict forecasts. How to evaluate the predictive power of the forecast systems will also be covered.

Speaker: Anders Kramer Knudsen

I'm a data scientist working at Cognitech. I have been working with data science fore more than 6 years. In that period, I have implemented multiple AI/ML solutions for customers in different industries. Several of them have won national as well as international credit. Before joining Cognitech I worked with complex business analysis within the financial sector.

Perfecting the last mile of the analytics supply chain

The target audience is data analytics and data leaders looking to understand and how to have meaningful data conversations with business people.

Talk with cover the concept of the analytics supply chain, review advances made in the last five years in terms of data engineering, and invite viewers to consider how the last mile of the supply chain - data delivery - needs to move beyond 'here's your dashboard, lmk if you have any questions.

Speaker: Ryan Dolley

Ryan Dolley is Head of Data at Count.co and a partner at analytics consulting firm PMsquare. His areas of expertise focus on business intelligence, data visualization and building data literate teams.

Forecasting at Scale – an Automated Approach

The presentation’s mainly addressing forecasting through the lens of AutoML. Additionally, some information on avantum & our approach to Automated Forecasting is given.

Target audience:

  • In a nutshell: anyone interested in AutoML for time series.
  • Or, more detailed: anyone interested in how to generate automated forecasts utilizing a broad selection of techniques ranging from neural nets over traditional methods like SARIMAX to tree based algorithms such as LightGBM. The perfect candidate would be someone looking to augment his or her decision-making capabilities, possibly having a finance background.

Prerequisites: No specific prerequisites though a basic familiarity with time series forecasting would be beneficial.

Take-aways: Automated ML for forecasting problems is not only possible but also a cost efficient yet effective way of bolstering decision making.

Speaker: David Trisl

For most of his professional and academic career David has been interested in distilling data and making subsequent insights accessible.

Focusing primarily on economic data analysis, he is currently working on making time series forecasting more accessible to everyday business users that stand to benefit from simplified access to state-of-the art forecasting solutions.

In his free time, he is trying his best to become proficient at Salsa Cubana even though having two left feet.

Why and When should I use IBM Business Analytics vs. PowerBI?

This session is about using the right tool to the right task. I will take a look at Planning and Cognos vs. PowerBI. You don't take your lawn mower to the the supermarket and you don't use a screwdriwer to the nails. We get a lot of questions about the differences between Cognos, Planning and PowerBI and I will give you an overview in this session without dissing any of the great products.

Speaker: Rikke Jacobsen

Rikke Jacobsen is the CEO & and Founder of CogniTech A/S. A Danish IBM Business Gold Partner. Keeping her employees happy and motivated and getting more people to use and get the most value out of IBM Analytics is her passion.

Rikke has worked with IBM Cognos Analytics for more than 20 years, and she still gets excited about the prospect of helping her clients make sense of their data. She is the first Dane to be named IBM Analytics Hero, a title that has been earned because of her enthusiasm and dedication to the Cognos community. A lot of people also know her as Mrs. Cognos.

Rikke has founded the Danish IBM Cognos and Planning Analytics User Group, where she inspires the users to get an even better value from Cognos. Rikke has educated over a thousand users in Cognos Analytics, she develops CogniTech’s own Cognos Courses and Cognos Training material. Rikke has inspired a lot of people as a speaker at several IBM Conferences and Webinars.

Rikke loves to travel the world together with her two adult children and husband, but a quiet day a home playing games with the family is also a favorite.

Tethering AI with Blockchain

Software architects, developers and technical managers will learn about the principles of designing and building AI that is tethered by blockchain. We will discuss why you want to be able to track and trace AI, and how you might use blockchain to reverse changes AI makes to itself.

Speaker: Karen Kilroy

Karen Kilroy, Kilroy Blockchain's CEO, is a life-long technologist with heart. Living in Austin, Texas, she has invented several products including CASEY, FLO, Kilroy Blockchain PaaS, RILEY, and CARNAK. Karen believes that artificial intelligence doesn't have to be scary and that technology can be proactively designed to be used for good. Karen's background includes working as an open source software engineer, which has prepared her for the new world of cognitive intelligence and tamper-evident ledgers. Named IBM Champion in the blockchain and AI categories, Karen is an experienced public speaker and author. She is the author of O'Reilly Media publications "Blockchain as a Service" (2019) and ""I and the Law" (2021), and the forthcoming "Blockchain-Tethered AI" (2022). Karen is also a professional dragon boat coach.

AI for Agave

Agave is one of the most emblematic crops of Mexico. Jalisco is home of the Tequila designation of origin, with only some municipalities in other parts of the country being also eligible to produce Tequila. Tequila production and agave cultivation are very important for the economic development of the state. In this context it is vital to develop strategies for monitoring the agave production. AI (Deep Learning) can conduct automatic monitoring of agave crops and plants.

Speaker: Eduardo Ulises Moya Sánchez

Ulises is the Director of Artificial Intelligence of the General Coordination of Innovation in the State of Jalisco, Mexico. Being the first director of this area in the public administration in Mexico.

He holds a Ph.D. in Electrical Engineering from CINVESTAV, a master's degree in Medical Physics from UNAM, and a Bachelor of Science in Physicist from the University of Guadalajara. He is a member of the National System of Researchers of CONACYT level 1. Last but not least, he is a founding partner of Nética, a company that seeks scientific dissemination for the training of young talents in STEAM.

He recently collaborated at the Barcelona Supercomputing Center in the high-performance artificial intelligence group, in deep learning projects for Retinopathies. In 2019, he was recognized with the Fulbright García-Robles grant to collaborate with the Quantitative Bioimaging Laboratory at the University of Texas in Dallas and the University of Texas Southwestern Medical Center.

Why is data quality important in an NLP project?

Topic: NLP data pre-processing

Audience: Data scientists, data engineers, students, anyone working with data. This talk might be a little too basic for some, or too advanced for others. Knowledge of basic ML techniques is useful, and even more useful is some knowledge of basic NLP.

Description: Data is everywhere, and in NLP data can come from any source: the web, your favorite book, linguist-compiled datasets, or even just your speech. When working with so much text data, it is important to clean up and normalize your text. This talk will cover the basics of data pre-processing in NLP, will introduce methods such as Word2Vec, and most importantly explain why NLP data requires pre-processing to be the most useful.

Speaker: Pierre-Nicolas Perrin

Pierre-Nicolas Perrin is a senior at UNC-CH studying computer science and mathematics, where he studied and performed ML/NLP research in classes on ML, NLP, Reinforcement Learning, Language and Learning, Optimization w/ focus on ML, and more. He interned at Capital One during the summer of 2021, and he is starting his Master's degree with a focus on ML and NLP in the fall of 2022.

Informix data analysis by Cloud Pak for Data

We will demonstrate how some of CloudPak for Data (CP4D) services can consume Informix data and how these services can be used for the simple analysis. Services like Watson Studio, SPSS or Data Refinery will be used as an examples for analyzing of the traditional relational data or TimeSeries data. We will focus on the analyzing of the Titanic passengers data and also native Time Series data processed first by Informix database. No prerequisites are required, the most part will be life demos and no PowerPoint.

Speaker: Jan Musil

Jan Musil is working 30 years in IBM as a IT specialist focusing on the database technologies, mainly Informix. Before entering to IBM he worked as a technical support engineer in Informix Software company.

He has also experiences with database security products like IBM Guardium, Test Data Management solutions like IBM Optim or database virtualization. His current main interest is in Kubernetes platforms, not only vanilla Kubernetes but also RedHat OpenShift. He is supporting Cloud Pak for Data solution and some micro services deployed there.

He loves walking or riding in the Czech magic nature with his wife, three kids and dog.

Is it time to switch to Linux on your desk?

There are substantial advancements in Linux desktop usability which may make it an excellent choice for you. Mike will share the secret sauce in how their team developed and delivered just-works Linux clients for professionals. This should be of interest to anyone that deploys to Linux. Modest prior experience with Linux environments is recommended. To find out more visit http://kfocus.org.

Speaker: Michael Mikowski

Michael Mikowski Best-selling author of Single Page Web Application: JavaScript end-to-end. Frequent conference presenter: FOSDAM, Chris Titus Tech, Manning.com Developer Tools and BigDaddyLinux. Web Application Architect, System Administrator, U.S. Patent-holder, Software consultant and designer. To find out more visit: https://michaelmikowski.com.

Detecting Icy Bridges with AWS, Image Processing and AI

My submission to the AWS Disaster Response hackathon, I'll use this git repo for the basis of my talk: https://github.com/rtp-aws/devpost_aws_disaster_response.

Speaker: John Fred Davis

John Fred Davis is an MSEE, Embedded/Systems Developer and Computing Enthusiast.

Logistics

  • AI on Pi Day is an online event.
  • Talks are in English.
  • Recording will be available online at a later date.
  • Attendance is free for AIDAUG members (and it is free to become an AIDAUG member).

Sponsoring opportunities

Contact Jean-Georges Perrin at jgp@aidaug.org. Do not hesitate to contact us.

(C) 2022, AIDAUG - Privacy

We are social