Fundamentals of Deep Learning Workshop for Multiple Data Types | SGInnovate
June 22
2018

Location

BASH, Level 3
79 Ayer Rajah Crescent
Singapore 139955

Fundamentals of Deep Learning Workshop for Multiple Data Types

Organised by SGInnovate, NVIDIA & NSCC

SGInnovate partners with NVIDIA Deep Learning Institute (DLI) to offer hands-on training to developers, data scientists, and researchers looking to solve real world problems with deep learning across diverse industries such as self-driving cars, healthcare, online services and robotics.

This full-day workshop explores how convolutional and recurrent neural networks can be combined to generate effective descriptions of content within images and video clips.

You will receive a Fundamentals certificate from Deep Learning Institute!

Attendees MUST bring their own laptops

 

Workshop Overview:
This hands-on course explores how convolutional and recurrent neural networks can be combined to generate effective descriptions of content within images and video clips.

You will learn how to train a network using TensorFlow and the MSCOCO dataset to generate captions from images and video by:

  • Implementing deep learning workflows like image segmentation and text generation
  • Comparing and contrasting data types, workflows, and frameworks
  • Combining computer vision and natural language processing

Upon completion, you’ll be able to solve deep learning problems that require multiple types of data inputs.

 

Recommended Prerequisites:

  • Taken/Attended “Fundamentals of Deep Learning for Computer Vision” or equivalent knowledge. Please refer to the following link for more info: https://nvidia.qwiklab.com/quests/11
     
  • TensorFlow and Python experience can be useful for some exercises
     
  • The mathematical and theoretical aspects of deep learning will NOT be covered by this training - and they're not a requirement to complete the labs, reading the Wikipedia page of Deep Learning would be a good start if you're interested

 

Agenda:

09:00 Registration
09:30 Image Segmentation with TensorFlow (hands-on lab)
11:00 Tea Break
11:15 Word Generation with TensorFlow (hands-on lab)
12:30 Lunch
13:30 Word Generation with TensorFlow (hands-on lab) [continued]
14:45 Image and Video Captioning by Combining CNNs and RNNs (hands-on lab)
15:45 Tea Break
16:00 Image and Video Captioning by Combining CNNs and RNNs (hands-on lab) [continued]
17:00 Closing Comments and Questions

 

Lab 1: Image Segmentation with TensorFlow
Image (or semantic) segmentation is the task of placing each pixel of an image into a specific class. In this lab, you’ll segment MRI images to measure parts of the heart by:

  • Comparing image segmentation with other computer vision problems
  • Experimenting with TensorFlow tools such as TensorBoard and the TensorFlow Python API
  • Learning to implement effective metrics for assessing model performance

Upon completion of this lab, you’ll be able to set up most computer vision workflows using deep learning.

 

Lab 2: Word Generation with TensorFlow
Predict the next word of a sentence using a Recurrent Neural Network. Neural networks can transform complex inputs to complex outputs with many different types of data. In this lab, you’ll train a network to predict the next word of a sentence using the MSCOCO dataset by:

  • Introducing Natural Language Processing (NLP) and Recurrent Neural Networks (RNNs)
  • Creating network inputs from text data
  • Testing with new data and iterating to improve performance

Upon completion of this lab, you’ll be able to train neural networks to understand both images and text.

 

Lab 3: Image and Video Captioning by Combining CNNs and RNNs
Learn to combine computer vision and natural language processing to describe scenes. Many applications of deep learning require the processing of multiple data types. Train a model that generates a description of an image from raw pixel data by:

  • Making use of the output of layers in the middle of neural networks
  • Combining data from multiple networks through concatenation and/or averaging
  • Harnessing the functionality of CNNs and RNNs

Upon completion of this lab, you’ll be able to combine workflows and data to innovate using deep learning.

 

Instructors’ Biodata:

Zhang Huaizheng:
Huaizheng is a second-year Ph.D. student at School of Computer Science and Engineering, Nanyang Technological University (NTU). His research interest is Deep Learning, Machine Learning, and Video Analysis. He has published two papers about applying Deep Learning to Video Analysis area on IEEE International Conference on Multimedia and Expo (ICME) 2018 and IEEE Transactions on Multimedia last year. He will also give an oral presentation at San Diego, USA this summer. He is good at PyTorch and Tensorflow and like sharing his Deep Learning code on GitHub. Besides research, he is also an experienced educator and has designed a course for Renaissance Engineering Programme at NTU, which is very popular, to teach students basic Deep Learning knowledge.

 

IMPORTANT PLEASE TAKE NOTE:

 

DLI Attendees Pre-Workshop Instructions

  1. You MUST bring your own laptop to this workshop.
  2. Create an account by going to https://nvlabs.qwiklab.com/ prior to getting to the workshop.
  3. MAKE SURE your laptop is set up prior to the workshop by following these steps:
    • Ensure websockets runs smoothly on your laptop by going to http://websocketstest.com/
    • Make sure that WebSockets work for you by seeing under Environment, WebSockets is supported and Data Receive, Send and Echo Test all check Yes under WebSockets (Port 80).
    • If there are issues with WebSockets, try updating your browser or trying a different browser. The labs will not run without WebSockets support.
    • Best browsers for the labs are Chrome, FireFox and Safari. The labs will run in IE but it is not an optimal experience.
  4. Please remember to sign in to https://nvlabs.qwiklab.com/ using the same email address as for event registration, since class access is given based on the event registration list.

 

For enquiries, please send an email to [email protected]

Upcoming Events

  • AI Readly

    This reading group is suited for AI researchers, academics and advanced practitioners. We meet every Tuesday, 7:00pm – 9:00pm at BASH, Level 3, Block 79 Ayer Rajah Crescent (part of JTC [email protected]-North) to discuss top recent AI-related arXiv preprint papers. 

  • Made in Singapore: Health And Life Sciences Edition

    Technology and innovation has great potential to be a great force in helping communities solve grand challenges in faster, more collaborative and data-driven ways. Join us in a panel discussion with innovators and entrepreneurs who are leaders in the area of health and life sciences. They will discuss the challenges they are solving - from drones to first responder vehicles to nurturing an ecosystem for engineers and computer scientists to build world-class tech companies from scratch.

  • Singapore Session: Collaboration for Innovation

    In celebration of the France-Singapore Year of Innovation, a signature Singapore Session conference will be held in conjunction with VivaTech in Paris, to share Singapore's innovation story as an AI and deep tech hub in Asia, and to deepen collaboration and foster partnerships with entrepreneurs, researchers and academics, investors, corporates and governments between French and Singapore innovation ecosystems. 

  • Python For Beginners

    Explore the intersection of coding and data with General Assembly. During our Python-focused introductory workshop, you’ll learn to harness the power of an essential programming language for data scientists.

  • Blockchain Programming for Programmers-Session 5

    This is a hands-on event - please come with laptops. We will be programming. At the end of the session, you should have working code. 

  • Made In Singapore: Aerospace Excellence

    Lauded as the inaugural “Aerospace City of the Future” by fDi (the intelligence portfolio of The Financial Times), Singapore boasts one of Asia’s largest and most diverse eco-systems with over 130 aerospace players, and is well positioned to build research activities, maintenance, repair and overhaul (MRO) services, and aircraft manufacturing.

  • Health Futures: Data Security and Privacy

    Data security is front and centre in the news every day. It is central to discussions on digital transformation of healthcare and the use of our most sensitive data: our health data. Panellists will discuss how technology systems are being used to protect patient data; how health data is accessed and used by payers, providers, and patients; and how regulations can enable or inhibit secure health data storage and sharing in Southeast Asia.

  • AI Readly

    This reading group is suited for AI researchers, academics and advanced practitioners. We meet every Tuesday, 7:00pm - 9:00pm at BASH, Level 3, Block 79 Ayer Rajah Crescent (part of JTC [email protected]-North) to discuss top recent AI-related arXiv preprint papers. 

  • Call for collaboration – Connect with Elsevier to innovate healthcare

    On 5 June, Elsevier is inviting you – medtech companies and startups to start a conversation with us. Join us to find out more about collaborations, partnerships and co-creation opportunities in healthcare. We will be offering our resources and expertise to collaborate with you in exciting medtech, healthtech projects in this region. 

  • User Data should belong to the User!

    With the new GDPR regulations said to take effect on 25th May 2018, it also set to create a big impact on how businesses all over the world handling user data and privacy. This event will discuss on the importance of data privacy and how users as well as businesses should have full control and access to their data. In the wake of the Facebook and Cambridge Analytica data misuse, public attention to the importance of data privacy has been heightened But the question that we need to ask ourselves is: “Do we necessarily have to wait for fiascos as such to be replayed on medias, to understand the intensity of having no access or control of our own data? When in fact, my data is mine. It should never be yours unless consented. Period.” For businesses this move will bring new opportunities never envisaged before. 

  • Talk Data To Me: Learning Resources and Books for Data Analysts and Scientists

    At this event, we host thought-leaders from the data community to discuss the possibilities that data brings to life. They’ll cover the best resources to start with data science and how to stay up to date.
    Why It Matters:According to a Forbes article, “2017 will be the year big data goes mainstream … whether gathering data on the front end or making big decisions in the C-Suite, every single person in your organization must buy into the value [data] analytics brings”.

  • Orange Fab Asia Pitch Competition in Singapore 2018

    Orange Fab Asia is holding a pitching event, supported by SGInnovate, in Singapore on 7 June to look for the next up and coming Asia-based startup with an established product. The 3 winners will be given the exciting opportunity to pitch and demo to large audiences from various VCs and large corporations during our Tokyo Demo Day held on 25 June. 

  • Morning Pitch Singapore: Robotics and Drones

    Morning Pitch is a pitching platform hosted by Deloitte Tohmatsu Venture Support. This platform helps start-ups form business alliances and partnerships with large corporations, corporate venture capital and venture capital firms. The focus on this Morning Pitch on 8 June 2018 is Robotics and Drones.

  • Blockchain Programming for Programmers – Session

    This is a hands-on event - please come with laptops. We will be programming. At the end of the session, you should have working code. At the session on 11 June, we will be going through more functions of ERC20, and building your own functions based on ERC20.

  • MedTech Meetup for Start-ups

    Bringing new medtech and digital health innovations to market is a complex and challenging endeavour. This informal event will bring together local SMEs and start-ups to discuss how to overcome some of these challenges, including raising capital. Join us at the APACMed-SGInnovate MedTech Meetup for Start-ups on Wednesday 13 June @ BASH (Ayer Rajah Blk 79).

  • Blockchain Evening: Accelerating Global Enterprise Blockchain Adoption

    The year 2017 was the year of experimentation as enterprises realized both the benefits and challenges of blockchain. 2018 will be a crucial stage for enterprises as they make a huge leap from proof-of-concept projects to full blockchain deployments. 

  • ScaleUp with SGInnovate-Microsoft Startup Pitching

    Introduction to Microsoft ScaleUp Programme: 

  • Design a T-shirt: Think Geek, Think Deep

    Passionate about STEM and want to be part of the rapidly growing science & technology community? Here is your chance! Whether you are a designer, coder or engineer by training, let your creative juices flow - YOU can show how cool deep technology is by designing a t-shirt that is uniquely coded.