Intel® Developers and Innovators were busy over the last month! Here’s an update on what the Intel® Software Innovators, Intel® Black Belt Software Developers, and Intel® Student Ambassadors were up to around the globe.
Peter MA demoed his Doctor Hazel project, which utilizes Intel® Movidius™ NCS and Intel® NUC. Justin Shenk demonstrated neural compute stick AI application, emotion detection, with a live camera-based application. Geeta Chauhan gave a talk and a webinar presentation of AI@Edge on Intel® Movidius™ NCS. Carlos Paradis presented a topic modelling application for cybersecurity on top of the Intel© Math Kernel Library. David Ojika gave a presentation on the usage of deep learning in high-energy physics domain as well as demoed deep learning at the edge using Intel© Movidius™ NCS and Intel Atom® based IoT® gateway in an application that did end-to-end recognition in videos from video camera. Eyal Gruss exhibited his Electronic Curator project as part of the NIPS 2017 creativity workshop.
BLACK BELT SPOTLIGHT
At IoT Day Catania Marco Dal Pino gave a session showcasing an IoT project that interacts with automobile sensors and interfaces that data to a cloud back end for predictive maintenance and also talked about the Intel® IoT Alliance workshops on the AWS cloud platform. Also at IoT Day Catania 2017 Massimo Bonanni hosted the Intel© IoT Alliance Workshop with AWS and also talked about Project Prague, an SDK created by Microsoft’s* Cognitive Services Labs that allows UX developers and designers to incorporate gesture-based controls into their apps.
Thomas Endres and Martin Foertsch gave a demo and talk showcasing their Avatar telepresence system using the Nao* Robot, Oculus* Rift, Intel® RealSense™ camera, and Intel© IoT Gateway as well as Genuino 101* technology. Abhishek Nandy published a book on Reinforcement Learning with Open AI and Tensorflow, co-authored with Manisha Biswas. Abhishek also spoke at two Commercial IoT Workshops and the Silly Meetup.
Rendra Toro spoke about and gave a training on how to use the Intel® Nervana™ platform and how to adopt it to your existing project at Seminar Nasional Machine Learning Conference, and also talked about how to use the IoT Gateway and connect it to your existing projects at GDG Bogor DevFest. Avirup Basu gave a hands-on workshop on IoT - and ML with Intel in Siliguri. Abhishek Sur gave a global office boot camp on cognitive data analytics and the integration of cognitive services with office products.
Adam Ardisasmita gave a hands-on lab on using Unity* library to create a VR game at Depok Margodev Meetup to help developers to code one time and port to all of the available VR platforms. At the Google Developer Festival in Bogor, Adam gave a talk on how to make VR games using Unity, highlighting the Intel® Core™ i7 processor.
Omkar Khair discussed getting started with ML and AI with Intel tools and Python distributions and also demonstrated the Intel® Deep Learning SDK for training models on Microsoft* Azure at Pune DevCon 2017. Siddhant Agarwal gave an overview of Intel IoT technology as well as how it integrates with Amazon Web Services at the Intel Alliance IoT workshop including neural networks, AWS educate programs, and demoes using the Intel® NUC gateway and AWS. He gave a similar talk, along with explaining the Intel® Movidius™ NCS and how it can used and the DevMesh platform at Google Developer Days Extended in Delhi.
As part of the World Usability Day event in India, Ghanashyam S gave a lightning talk about “How to choose the right UX Metrics” in the mobile and web design area at Google Developer Days India. He also was one of the mentors at the events Office Hours answering questions and giving advice regarding design sprints for students and developers as well as facilitated a three-hour Design Thinking workshop for professionals and students. Rishabh Banga led a Learn to Lead – Getting Started with IoT and AWS webinar which gave an overview of Intel IoT technology as well as how it integrates with Amazon Web Services’ new Amazon* Rekognition API and a demo of the developer kit. Nishant Srivastava shared his project, Sensey, which is an Android library that makes playing with sensor events and detecting gestures a breeze and his project, Validate Tor, which is an Android library for fast and simple string validation.
Pooja Baraskar gave a session on hardware and software knowledge with an in-depth understanding on how to work with the Intel® Commercial IoT developer kit to help start the product development with a fresh approach and the latest available technologies in the market at Techbhubaneshwar Conference. Manisha Biswas co-authored a book with Abhishek Nandy that teaches reinforcement learning, a popular area of machine learning, starting with the basics, working through theories, and building up the concepts of the reinforcement learning process. She also organized two Intel Alliance IoT workshops, and spoke on commercial IoT with Amazon AWS at the Rise of Women Entrepreneurs meetup which has a focus on empowering women with technology.
Cristina Dominicini shared a paper on her VirtPhy project which is a fully programmable NFV orchestration architecture for edge data centers that is designed to tackle the demands of edge computing such as throughput, latency, cost, innovation, and efficient orchestration required in the geographical distribution of small-scale data centers. Nelson Glauber spoke about modern Android application development techniques at DevFest XP in Brazil. Pedro Kayatt gave a SharpTalk on virtual reality, augmented reality, and mixed reality at Pixel Show 2017.
Rosemarie Day hosted two IoT workshops at her old technical high school where she talked about open source technology and the Intel Innovator program. Ritik Patnaik developed Picrosccope which is an affordable, DIY, RPi-powered microscope that allows you to create and interact with your very own micro-world. It’s a great hands-on project for someone who is interested in bio-technology and the worlds of microbiology, optics, or DIY electronics. It can be built within a day by nearly anyone, regardless of age or skill-level, with a few electronic components, a 3D printer, and a $60 budget. Jesse Banks held his regular lab night as well as had a Virror build party which worked on the construction of a large interactive art piece that will be showing at PXD WLF in February.
Alex Porter wrote three blog posts recapping Austin, Texas meetups and events that Underminer Studios was a part of including: AR/VR Tools & Tech December Meetup, Video Game Makers Unite December Meetup, and DeveloperWeek Austin 2017. Tim Porter participated in the Hoft Institute Portfolio Review which was a workshop that had 1:1 sessions to help critique and educate students about career possibilities and how to improve their skills and portfolios for the job market. Peter Ma shares a video of his Clean Water AI which uses AI and microscope camera to measure water safety by detecting dangerous bacteria and harmful particles in the water source. He uses the Intel® Movidius™ NCS to detect water contamination, update the map, and give users information about water safety in real time.
Harsh Verma held a planning session for the ACM’s upcoming online virtual education sessions as well as planning out activities like AI and Machine Learning sessions using the Intel® Movidius™ NCS and VR on Intel platforms for the upcoming year.
Nader Rahman gave a training on developing a dashboard for the supervision of an Intel® Galileo card-based sensor network and node JS server at ESPRIT University. Alejandro Alcade wrote the second post of his Crpytography 101 course, which teaches a user how to computer modular exponentiation. Eskil Steenberg has been streaming programming in the area of graphics throughout December and answering questions and helping people get started with game development.
Maksim Masalski demoed his self-driving robotic cars with Intel RealSense cameras which is intended to be part of the educational material which helps students to study how to program self-driving cars in high schools and universities at Maker Faire Rome 2017. Eyal Gruss exhibited his Electronic Curator project as part of the NIPS 2017 creativity workshop and at Utopia Festival in Tel Aviv. Eyal also gave a shallow introduction to deep learning as part of the ITWorks conference to advance Arab students integration into high-tech at Ben Gurion University, Musrara Art School, and Tel Aviv Academic College.
Fabrizio Lapiello shares his BTM project which monitors the vibrations and even the minimum amount of shaft movement among big trees. He gave an Intel IoT Alliance Workshop at the University of Naples. Fabrizio also wrote three articles as part of an IoT 101 Series on Designing IoT Devices that he is working on: Design IoT Devices Introduction, Technologies Design, and Usability Affordance. Michael Schloh gave an interactive Intel AWS Optical Computer Vision workshop at RIAT which focused on OpenCV and AWS IoT. Johnny Chan was interviewed for an article on IDZ, as well as wrote a tutorial on training a basic wild mushroom classifier. Pascal Van Kooten presented his work and comparison of frameworks regarding automated machine learning at AI Fyraftensbajer CPH meetup.
Oluwatobi Oyinlola gave a talk on why build apps for IoT with Python* at that Python Ibadan meetup in Nigeria. Paolo Moro presented his team’s Wingsuit VR Vitruvian Game project at Maker Faire Rome 2017 and were recognized as having one of the best stands of the fair and about 300 people were able to successfully test their prototype. Ron Evans had his recent talk “Putting Eyes on the IoT” from the Global IoT DevFest 2 featured in Hackster’s newsletter which was sent out to over 300k recipients. Liang Wang spoke at the Bloomberg Data Science Meetup in London. Zayen Chagra wrote a LinkedIn article (in French) for getting started in Commercial IoT.
Salvino Fidacaro taught an Android* Apps Builder course for Codelab School which was a free activity for reserved students and walked through the development of an android application oriented to the world of VR. Justin Shenk demonstrated neural compute stick AI application, emotion detection, with a live camera-based application at NIPS 2017. Justin also gave an introduction to developing neural networks and edge computing with Intel® Movidius™ NCS for mobile EEG at Institute of Cognitive Sciences Winter Seminar. He also delivered a presentation on “Visualizing neural networks with open source projects” at the Lower Saxony AI/Machine Learning Meetup.
STUDENT AMBASSADOR SPOTLIGHT
Anirudh Bhardwaj gave a workshop on machine learning and deep learning at Vellore Institute of Technology Chennai Campus focusing on an introduction to machine learning, deep learning, and applications of each, Intel© Optimized libraries, Caffe, and other open source resources, Intel© Optimized MKL, and Intel© Optimized Toolkits. Prajjwal Bhargava shared two projects, Neural Style Transfer, which is a Tensorflow implementation of a neural algorithm of artistic style and makes use of convolutional neural networks to combine the features of two images; and Wasserstein GANs, which is a repository contains implementation of generative models. Avinash Madasu shared his project which is a sentiment analysis of movie reviews.
Pallab Paul hosted an Intel AI workshop at Rutgers University where they used the Intel® Nervana™ Deep Learning kit provided by Intel to teach fellow students about the use cases of deep learning and how to get started. Tejeswar Tadi and Pallab Paul were featured in an IDZ interview blog post about the Future Labs AI Summit in NYU.
At the OSU Machine Learning Club Meetup Christian Gabor discussed Intel’s Machine Learning 101 and Deep Learning 101 courses using Jupyter* Notebook as well as future plans and projects for the club including projects in the financial market prediction, self-driving RC cars, and Starcraft*/game AI areas. Aman Mathur shared his project, Travel Assist Ontology, which uses sentiment analysis of reviews to assist the user in choosing the accommodations, flights, and restaurants for their travels that is the best fit for them.
Soheil Jamshidi gave a lecture on IoT vulnerabilities reviewing the basics of definitions, usage, and application development of IoT devices along with known vulnerabilities of IoT devices both on system design side and also developer/end user side by focusing on IoT botnets such as Mirai*. The goal was to make sure that students who are interested in using IoT devices such as Intel® Galileo are aware of issues and risks associated with careless usage of them. Dipanjan Dutta shared his project which intends to develop an efficient machine learning algorithm to be able to predict the 11 on-field starting players for an upcoming association football match.
Alfred Ongere, with the help of Derrick Mungai and Ruth Waignanjo, gave a hands-on workshop at the Technical University of Mombasa on the fundamentals of machine learning while exploring examples in natural language processing, navigation, reason and perception manifestations of AI. In addition they spoke on AI best practices and took the attendees through a hands-on session on constructing a simple binary classifier. Alfred also gave another hands-on workshop at the University of Eldoret, with the help of Gilbert Kigen, where they took the attendees through the fundamentals of AI using real-world examples, gave the hands-on session on binary classification, and also introduced deep learning while using neural network playground by Tensorflow.
As part of his Senior Project Final Demonstrations Damla Gul Altunay presented his work on abandoned luggage detection using Tensorflow and OpenCV at Middle East Technical University. Abraham Ogol put on a workshop at Maseno University; it was a continuation of an earlier workshop and focused on how to get started in AI, including ML and neural networks. The workshop also included a presentation on Android* Things and Tensorflow* and offered a variety of demos including vision kit, data visualization and modeling, and TensorFlow integration into Android Things.
At Yantra 6.0 Kshitiz Rimal gave a talk on courses and software provided by Intel for machine learning including a brief introduction to Intel® AI DevCloud and some insight on getting started with deep learning as well as hosting a hackathon at the event. Karandeep Singh Dhillon has applied deep learning and computer vision for Amazon Basin Land Classification and also used Keras and further ready to run code using Intel© Movidius™ NCS in his Human Footprint Classification project. Suprabhat Das shared his project which is designed to apply deep learning and artificial intelligence in making crucial rocket parts to bring about revolution.
Chris Barsolai gave Intel AI deep learning workshops at both Maseno University and Dedan Kimathi University of Technology. Emanuel Di Nardo created a new blog to show people what machine learning and AI are, with an aim to divulgate AI to a great number of people, both experts and novices, in order to create greater understanding of this field. Eren Halici gave a hands-on workshop on Deep Learning Tutorials at Middle East Technical University.
Vincenzo Santopietro wrote a blog post giving a brief overview about the Intel® Movidius™ Neural Compute Stick device. Felix Farley and his team, SuperPlug, came in third place in the Porticode 2.0 hackathon in London which was run by the UCL Technology Society and focused on IOTA blockchain implementation and API endpoints. Szymon Kocot had a meeting with the medical faculty and doctors about their ongoing project researching Parkinson disorder biomarkers using convolutional neural networks and class activation maps from T1, T2 MRI imaging, showing their methods and what they’ve discovered so far. Yaron Michl gave a lecture on the use of Google* Earth Engine with the Python* API at the Israeli Geographical Association 2017 conference.
N. Aravindham participated in the Gains AI Hackathon in Hyderabad and the Pragyan Hackathon in Bangalore. Salil Gautam shares his EvnCoin project which is an Ethereum* based blockchain was made with the motivation of reducing the amount of pollution in the air we breathe, and combat pollution in a general sense. Shubham Jaiswal shared his project which is a conversation agent for real estate users; essentially an automated grievance system Skype-bot for real estate users using the Microsoft* Bot Framework. Viabhav Amit Patel shares his project for solving graph isomorphism problems for special cases.
Daniel Theisges dos Santos shared his Brazilian Traffic Signs project which is an advanced driver assistance system designed to recognize typical Brazilian traffic signs through computer vision techniques, algorithms to filter objects, and using a deep neural network for isolating objects. Also, unrestrained handwritten text recognition is a very challenging problem and Daniel presents a model composed of a hybrid of convolutional and recurrent neural networks that recognize handwritten text using the image pixels as input features based on deep learning techniques to solve it.