Frontiers of Data Driven AI
Professor Hui Xiong, The Head of Business Intelligence Lab at Baidu Inc.
Abstract: In this talk, I will present some emerging opportunities and challenges for data driven AI with an eye on trends in deep learning, recommendation engines, and data-intensive computing platforms, as well as applications such as autonomous vehicles, B2B marketing, location-aware social media, car insurance pricing, and talent analytics.
Building cyber threat analyst centaurs using AI and machine learning
Staffan Truvé, CTO & Co-founder, Recorded Future
Abstract: This talk describes how recorded uses AI and machine learning to do information harvesting and analysis in real time and at web scale, to provide threat intelligence and predictive analytics to the world’s most demanding customers.
How Artificial Intelligence shapes a smarter society in the Nordics
Christian Guttmann, Vice President and Global Head of Artificial Intelligence and Data Science, Tieto
Abstract: According to Tieto’s 2017 survey almost three out of four (71%) people in the Nordics believe that new technology, driven by AI, will help make their everyday life easier. Not only the citizens believe in this: as an example, Vinnova – the Swedish innovation agency, financed 142 AI projects in 2017, to be compared with the 13 AI projects in 2013.
In this presentation Christian Guttmann share insights about the Nordic marketplace, and examples of how we extend and apply AI technologies specifically into the healthcare & welfare sector. Christian presents how Tieto addresses the societal and industrial challenges deeply entrepreneurial, through collaborations with leading scientific institutes and industry partners, as well as opportunities offered to AI talent. One example is how Tieto helped the city of Espoo in Finland to analyze social and health data of the entire population of Espoo to identify citizens that are at risk of social exclusion. The trial showed that our AI technology picked up risk factors that trigger need for social support, and thus has the potential for a substantial reduction of the economic and social burden that results from social exclusion.
From AI to Precision Medicine: Epileptic Seizure Prediction Using Big Data and Deep-Learning
Stefan Harrer, Manager Brain-Inspired Computing, IBM Research – Australia
Abstract: We have achieved a world-first breakthrough in mobile, personalized, epileptic seizure forecasting with AI. The main creative elements are the use of Deep Learning for decoding brain states, the use of a custom developed neuromorphic processor allowing to deploy such models on a wearable device, and the introduction of tunable algorithms allowing the patient to set desired alert modalities. I will explain IBM’s approach and positioning in AI for Healthcare with a special focus on our latest work on AI-driven epilepsy management and treatment technology.
Voice Assistants：Next Generation of Human-Machine Interaction
Zaiqing Nie, Researcher and Senior Director, Alibaba AI Labs
Abstract: Voice assistants are becoming part of our everyday life, and lots of daily user scenarios could be enabled through voice interaction. In this talk, I will discuss the challenges in building intelligent voice assistants and introduce our conversational AI platform for developers and data scientists to easily connect their contents, services, and devices to voice assistants, with maximum productivity.
Using conversational AI for better customer experience
Hui Wang, Senior scientist, Xiaoi
Abstract: The popularization of Internet and mobile communications has created huge convenience for customers. However, it also set a higher requirement for customer service. In our session, we will introduce how our conversational system help our clients improve customer experience while reduce operation and labor cost. : In China Construction Bank, the system has provided services for over 1 billion customers within three years. China Merchants Bank uses only 10 staff to handle all requests with the help of the system. The timely, fast and accurate response of AI system creates brand new customer experience. In the process of deep integration with the clients, Xiaoi evolved from basic AI to deep AI, and provides more contextual interactions to respond to customers’ request in a more timely manner.
Augmented Reality: Real-Time 3D Localization and Video Object Segmentation
Wei Liu, Distinguished Scientist and Director, Tencent AI lab
Abstract: This talk mainly introduces our recent advances in two vision topics related to augmented reality: real-time 3d localization and video object segmentation. In real-time 3d localization, two novel algorithms will be introduced: a lightweight anchor-image based real-time 3d localization algorithm specially designed for augmented reality, and a visual-inertial odometry algorithm that models varying camera-imu time offsets. In video object segmentation, a novel algorithm, namely CNN-in-MRF, which embeds a CNN as a spatial energy in a spatio-temporal MRF to solve the segmentation problem, will be introduced.
Building a hybrid intelligent system for personalized digital sports coaching
Dr. Mohammed Adel El-Beltagy, CTO RaceFox
Abstract: At RaceFox we help athletes better understand their technique and achieve their sports goals in an efficient and safe way. To realize this, we start with motion classification from sensor data and then extract human understandable KPIs that allows for greater sense making. In order to address specific motion limitations and to plan exercise sessions, we illicit expert knowledge to power our coaching system. We combine various AI techniques to give our athletes a seamless coaching experience. In this talk we will present our AI coaching architecture and we will delve into the integration between different AI subsystems.
AI-Powered Smart Retail at JD.com
Jinfeng Yi, Director of Machine Learning Lab, JD AI Research
Abstract: As China’s largest retailer, JD.com is constantly raising the bar for the customer experience, from product selection and recommendation to delivery logistics. At the heart of this – from understanding how customers behave and predicting trends in demand to developing sophisticated speech recognition technology that allows customers to shop online using voice commands alone – is JD’s sophisticated AI technology. JD will provide insights on how the company is using machine learning, computer vision and natural language processing to engage consumers in entirely new ways and chart the future of retail.
Machine Learning and AI Research for Digital Marketing
Georgios Theocharous, Senior Research Scientist, Adobe
Abstract: Adobe is a company well known for its creative and multimedia software such as Photoshop, PDF and Flash. Lesser known is the fact that Adobe innovation can be experienced everywhere. Since 2009, Adobe has grown into a powerhouse of digital experience marketing and with big data analytics solutions handling hundreds of trillions of transactions annually and leveraging PBs of data from the top web and mobile brands. In this talk, I will describe Adobe’s Experience Cloud, which powers the various digital marketing experiences. I will talk about Adobe Research, which addresses many of the AI and machine learning problems for digital marketing. Finally, I will summarize some of the research challenges pursued in my group around the area of decision making.
AI for Transportation
Yan Liu, Chief Scientist, DiDi AI Labs
Abstract: This talk is about how AI technologies have been applied to analyze such big transportation data to improve the travel experience for millions of people in China.
Autonomous driving and industrializing of AI at scale
Nicholas Wickström, Product Owner Vision/Deep learning for Cruising and Highway Autonomous Driving
Abstract: An introduction to autonomous driving and the challenges involved in the industrialization of computer vision based on machine learning.
From Theory to Practice: Applying AI to Enterprise Applications
Anand S. Rao, Global Artificial Intelligence Lead, PwC
Abstract: The adoption of AI in the consumer sector has been taking place for over a decade, while the use of AI in enterprise applications is just beginning. Most enterprises are in the early stage of AI adoption working on their AI and Automation strategies or developing proof-of-concepts to assess the overall business impact of AI. Some leading enterprises are moving into production AI and starting to think about ‘responsible AI’ that addresses issues around bias, transparency, explainability etc. In this talk, we show how AI is being used by enterprises across all sectors, across different elements of their respective value chains. We provide use cases along multiple dimensions – sectors, functional areas, technologies (e.g., machine learning, NLP etc.), and data types. Based on the lessons learnt from our client and innovation work we outline the roadmap for enterprise clients.
Diversity and Depth: Implementing AI across many long tail domains
Paul Groth, Disruptive Technology Director, Elsevier
Abstract: Elsevier serves researchers, doctors, and nurses. They have come to expect the same AI based services that they use in everyday life in their work environment, e.g.: recommendations, answer driven search, and summarized information. However, providing these sorts of services over the plethora of low resource domains that characterize science and medicine is a challenging proposition. (For example, most of the shelf NLP components are trained on newspaper corpora and exhibit much worse performance on scientific text). Furthermore, the level of precision expected in these domains is quite high. In this talk, we overview our efforts to overcome this challenge through the application of four techniques: 1) unsupervised learning; 2) leveraging of highly skilled but low volume expert annotators; 2) designing annotation tasks for non-experts in expert domains; and 4) transfer learning. We conclude with a series of open issues for the AI community stemming from our experience.
Machine Intelligence at Ericsson
Elena Fersman, Research Director, Machine Intelligence and Automation, Ericsson
Abstract: At Ericsson we use Machine Learning and other AI technologies to automate our products and services. In 5G, AI technologies are an integral part of making the networks meet the new challenges and achieve performance that far exceeds network performance of today. In order to create value from data in real time we use methods such as reinforcement learning, deep learning, semantic reasoning and planning. By using these technologies we are able to maximize the output for each function in our highly distributed and decentralized networks making them more resilient, adaptive and optimized than ever possible before. Machine Learning and other AI technologies also allow us to fuel new innovation we can’t imagine today.
Adaptive Learning Powered by AI – From Model to Implementation
Abstract: Songshu AI by Yixue Education combines the advanced AI technology with custom-build and calibrated learning contents to offer after-school tutoring and supplementary academic programs, through both online and connected learning centers, to K-12 students in China. In this talk, Joleen Liang, Partner of Songshu AI Learning and Richard Tong, the chief architect of Songshu AI will offer an in-depth review of the AI models, the adaptive system architecture, and pedagogical model behind Songshu AI and shed lights on the unique challenges and lesson learned from our adaptive learning implementation in the after-school market in China.
Joleen Liang, Partner of Songshu AI Learning
Richard Tong, the Chief Architect of Songshu AI and General Manager of Songshu AI Learning US Operations