Artificial Intelligence (AI) has been a hot topic for several years now, constantly finding its way into a growing number of everyday applications.
In this track, you'll have access to the new trends in AI and the opportunity to learn more about its various subfields, such as machine learning, deep learning, natural language processing (NLP) and robotics.
Wednesday, June 9, 2021
09h às 19h GMT-3
REMOTE ACCESS WITH ONLINE BROADCAST
For Brazilians, in BRL:
1 track: R$ 145 for R$ 110
2 tracks: R$ 290 for R$ 198
3 tracks: R$ 435 for R$ 285
* price valid until APR/26,
see full table
For Brazilians, in BRL:
1 track: R$ 145 for R$ 130
2 tracks: R$ 290 for R$ 230
3 tracks: R$ 435 for R$ 330
* price valid until MAY/28,
see full table
For Brazilians, in BRL:
1 track: R$ 145
2 tracks: R$ 290 for R$ 260
3 tracks: R$ 435 for R$ 370
* price valid until JUN/10,
see full table
For Foreigners, in USD:
1 track: $30 for $20 USD
Connect Pass: $80 for $60 USD
* price valid until MAY/08
For Foreigners, in USD:
1 track: $30 for $25 USD
Connect Pass: $80 for $70 USD
* price valid until JUN/01
For Foreigners, in USD:
1 track: $30 USD
Connect Pass: $60 USD
* price valid until JUN/10
Time | Content |
---|---|
10:15 to 10:45
(GMT-3) 13:15 to 13:45 (GMT) |
TDC OpeningAt International Stadium with Yara Senger
|
10:50 to 11:25
(GMT-3) 13:50 to 14:25 (GMT) |
Machine/Deep Learning for Accurate Diagnosis of Covid-19Noureddin SadawiIn this presentation, I am going to present and review three techniques that employ machine/deep learning for the accurate diagnosis of Covid-19. These techniques are powerful and can save time and effort. Their performance is comparable with that of polymerase chain reaction (PCR) test. In fact, they can be better than human expert diagnosis. All of this and more will be covered in the talk. |
11:30 to 12:05
(GMT-3) 14:30 to 15:05 (GMT) |
Deep Learning-Driven Text Summarization & Explainability for Legal DocumentsNina HristozovaLegal documents tend to be very long and complex, therefore summarizing them is a laborious and time consuming task. Thomson Reuters Labs added innovative Deep Learning (DL) capabilities to an existing editorial tool to augment this experience. A DL-powered summarisation model was built to automatically generate an initial version of those summaries, which are then verified by internal editors. Having a machine-generated summary helped our editors a lot in terms of time savings. But, is there a way to add even more value for our editors and increase their trust in this AI system? To answer this question we turned to the concept of Explainable AI. Join our talk to find out HOW! |
12:10 to 12:45
(GMT-3) 15:10 to 15:45 (GMT) |
Deep Learning for Tabular Data using PyTorch TabularManu JosephPyTorch Tabular - a new deep learning library which makes working with Deep Learning and tabular data easy. Built on top of PyTorch and PyTorch Lightning and works on pandas dataframes directly. Many SOTA models like NODE and TabNet are already implemented in the library with a unified API. PyTorch Tabular is designed to be easily extensible for researchers, simple for practitioners, and robust in industrial deployments. Talk Outline
|
12:50 to 13:50
(GMT-3) 15:50 to 16:50 (GMT) |
Networking and Visiting Stands
Break to network and get to know the booths of the event. |
14:00 to 14:05
(GMT-3) 17:00 to 17:05 (GMT) |
Track opening by coordination
Here the coordinators introduce themselves and make an introduction to the track. |
14:10 to 14:45
(GMT-3) 17:10 to 17:45 (GMT) |
Yearn to LearnMatthew EmerickDespite a skills shortage, there are millions of people interested in artificial intelligence. How do you get from zero to hired? What path should you take? How do you navigate all of the available resources? Are you stuck in tutorial hell? How do you stand out? Join me as we discuss the opportunities, both technical and non-technical, for anyone and everyone interested in such an amazing field. Be part of the future with this tour de force of information that can help you move forward. |
14:50 to 16:05
(GMT-3) 17:50 to 19:05 (GMT) |
How to introduce AI into my company's products?Flávio Henrique Moura Stakoviak / Ricardo Fernandes / Matthew EmerickPainel Digital desta Trilha \ Digital Panel of this Track
|
16:10 to 16:25
(GMT-3) 19:10 to 19:25 (GMT) |
Networking and Visiting Stands
Break to network and get to know the booths of the event. |
16:25 to 17:00
(GMT-3) 19:25 to 20:00 (GMT) |
Responsible AIPedro LelisLet's start by introducing some cases in order to exemplify the importance of responsibility when working with AI systems. We'll talk about survivorship bias, loan decisions for different groups and bias / variance trade off. Finally, we'll discuss good practices to promote fairness, interpretability, privacy and security. |
17:05 to 17:40
(GMT-3) 20:05 to 20:40 (GMT) |
Generating Classic Nintendo Music with Python and MagentaSamuel AgnewArtificial Creativity has been a topic of debate among scholars, artists, and computing experts. What's not controversial is that writing code to generate music is a ton of fun. Tools like Magenta, a Python library powered by Tensorflow for handling music data, make it easier for developers to use machine learning for music without being neural network experts. In this talk we will walk through how to create music with Magenta using a recurrent neural network trained on the soundtracks from classic games on the Nintendo Entertainment System. We will also live-code a web application which uses the Twilio API to provide a phone number people can call to listen to computer-generated music. |
17:45 to 18:20
(GMT-3) 20:45 to 21:20 (GMT) |
![]() Understanding Machine Learning for Java DevelopersFrank GrecoMachine Learning (ML) is a huge, long-term global trend that affects every part of the stack from the user to the hardware. Visual Recognition (VisRec) is an important subclass of ML with wide business applications across many types of industries and use cases. But for Java developers, there doesn?t seem to be many ML coding options other than learning Python. The current ML libraries either are very complex and designed for data scientists, or they are Java wrappers around C/C++ libraries and don?t ?feel? like Java tools. JSR381 is a new Java-native API for ML. We?ll talk about goals with JSR381, the API, its internal architecture, and show some examples. |
18:25 to 18:45
(GMT-3) 21:25 to 21:45 (GMT) |
Closing session
After the presentation of the results of the day, on the Stadium stage, many sweepstakes will close the day. |