In our last Global Machine Intelligence Summit 2017 preview, we introduced the technical sessions, delving into the current research topics of Jürgen Schmidhuber, Stuart Russell, Gary Marcus, Martin Müller and other speakers.
Now, with just one day remaining before the GMIS 2017 kicks off in Beijing, we’ll survey the industry applications of AI technologies. GMIS 2017 has invited Xiaochuan Wang, CEO of Sogou; Weiying Ma, Director of toutiao.com Lab; Dong Yu, Deputy Director of Tencent AI Lab; Yu Hu, CEO of iFlytek; Kai Yu, Founder of Horizon Robotics; Wenyuan Dai, Founder of 4Paradigm; Qiong Yang, CTO of Paiyipai; and Turing Robot Founder Zhichen Yu to share their views on the current AI application landscape and update us on their latest practices.
Speech Dialogue Application
Tencent: AI Business Strategy
The Tencent AI Lab was founded in 2016, with research focus on computer vision, speech cognition, natural language processing, and machine learning. For business applications, Tencent incorporates its own competitive advantage as China’s largest social media platform to make strategic deployment in gaming, social, content generation, and platform tools. Gaming improves decision-making systems and reinforcement learning algorithms; content generation promotes personalized and targeted media/advertisement content distribution; while AI platforms aim to provide open solutions via cloud services to build a deep learning ecosystem.
Voice recognition is one of the most important technologies in human-computer interaction (HCI) and a top focus at the GMIS 2017. Dong Yu, Deputy Director of the Tencent Artificial Intelligence Lab, former Chief Researcher at Microsoft’s Speech & Dialog Research Group will brief us on the frontier development of this technology.
AISpeech: Working Alongside Chip Manufacturers
AISpeech focuses on human-computer voice interaction and has decades of technical built-up in this field. At the moment the company’s voice technology allows real-time voice recognition, contextual semantic understanding, correction, interruption, and multi-rounded dialogues. The company is currently working alongside chip manufacturers to integrate speech recognition technology with hardware. Kai Yu, Co-Founder and Chief Scientist at AISpeech, will deliver a keynote session on moving towards intelligent cognitive dialogue interaction.
iFlytek: Outcompeting Google in Mandarin Speech Dialogue Technology
iFlytek is among the most prominent AI speech dialogue companies today, outcompeting Google in Mandarin voice recognition technology. The company has invested heavily in the “iFlytek Brain” project since 2016, to improve machines’ listening, speaking, and comprehension. iFlytek’s current revenue stream is ¥33 billion CNY, an increase of 32.78% over last year.
In addition, iFlytek is working to integrate voice recognition technology in medical and legal areas. Doctors will be able to write medical records through voice inputs and quicken the process of diagnosis. CEO Yu Hu will speak on real-world applications for artificial intelligence.
Natural Language Processing Application
Toutiao’s Bet on AI produced Media Content
During the 2016 Rio Olympic Games, an AI reporter allegedly wrote 457 newsletters on badminton, table tennis and tennis tournaments, generating 30 pieces a day. This makes the life of flesh-and-blood journalists difficult as the tireless Xiaoming Bot uses natural language processing, visual graphics processing and machine learning technologies to perform the task.
In addition to Xiaoming Bot, Toutiao also set up an AI Lab to optimize content recommendation by labelling topic relevance, context, trend, and producer. The process comprises NLP, deep learning, smart Q&A systems, computer vision, big data analytics and so forth to help optimize content production and distribution.
Dr. Weiying Ma, Director of Toutiao AI Lab and Toutiao Vice President, will deliver his keynote speech on “The Future of Information Flows and their Opportunities with Artificial Intelligence.”
Sougou’s “Wangzai” (Woofie) Bot Challenges Humans
Sougou’s Q&A bot Wangzai recently found itself in the media spotlight for outcompeting human opponents on TV shows. Sougou started off as a search engine and input method company and currently focuses on the R&D of deep learning, machine translation, voice dialogue, and intelligent Q&As. Company CEO Xiaochuan Wang will join a panel on the insights and reflection of AI, and demonstrate Wangzai’s voice recognition talents on-site at the GMIS with a competition against humans.
Emotibot: Make the Machine Feel Us
Former Microsoft (Asia) Internet Engineer Institute Vice President Renxian Jen founded Emotibot to help synthesize emotions with dialogue systems.
Emotibot converses with users and generates reciprocating emotional responses. It can now identify over 200 kinds of human intentions, initiating task-oriented, chat-oriented, and information-oriented dialogues. It can also help the machine discern over 22 kinds of human emotions, and build a multi-modal emotional recognition model to help the machine listen, read, and visually comprehend what is going on. Jen will guide us through the details of this application in his keynote speech.
Turing Robotics: Make the Machine Understand the World
According to Zhichen Yu, Turing Robotics emphasises the importance of “making the machine understand the world.” The company’s core technology is based on natural language processing and semantics, and provides a great service platform for developers and partners. Its current products involve open access personalized intelligent robot platforms and OS operating systems used by robot manufacturers.
In July 2016, 4Paradigm released “Prophet,” an AI platform based on deep learning, reinforcement learning and transfer learning. By setting parameters on automated algorithms, laying out infrastructure for machine learning, providing automated or semi-automated feature engineering, model selection, and reducing reliance of data scientists, the current public cloud version of Prophet can provide solutions for financial, medical, media, and electrical applications. Paradigm’s Founder Wenyuan Dai4 has led Baidu’s search marketing team using the Phoenix Nest system. The system implements deep learning and has helped enhance Baidu’s liquidity by eight times over the past four years. Dai will brief us on the commercialization of AI and its unlimited value.
Horizon Robotics: Autonomous Driving Platforms
Horizon Robotics’ “Brain” platform focuses on deep neural networks for software and hardware. Products embedded with an AI brain will be able to perceive their surroundings, make decisions and generate thoughts.
Horizon has launched operating systems based on deep neural networks, product-inlaced “Hugo” for autonomous driving, and “Andersen” for home appliances. The former can detect road conditions, and make predictions based on interacting traffic agents; while the latter can detect human body temperature and adjust AC settings accordingly. Horizon is also working with chips to implement AI onto hardware. CEO Kai Yu will lead us further into the discussion with his keynote speech on “Deep Driving in Automated Driving.”
Finance, Healthcare, Cognitive Intelligence, and Operations Research
Alibaba’s Fintech: Ant Financial Services
Alibaba’s AI development focuses on business service application scenarios. Yuan Qi, Chief Data Scientist and Vice President at Alibaba’s Ant Financial Services leads his team in implementation of the technology on all sectors and service facades. The company has made significant progress on intelligent customer service and smart marketing. His keynote will help us understand the tremendous value of AI technology in financial applications.
PaiYiPai: Applying AI Technology in Healthcare with Products that Work
Everyone is aware of the huge potential of AI in healthcare. PaiYiPai uses optical character recognition (OCR) technology to help machines read medical documents. The system currently supports blood, urine, tumour, coagulation, viruses, and hundreds of other common lab tests; it can also identify medical invoices, prescriptions, diagnostic certificates, sort different documents, and perform data mining.
PaiYiPai can also help analyze medical images including CT, X-ray, MRI, PET, pathological section, fundus photography, ultrasound and so on. For X-ray images, PaiYiPai uses convolution neural networks to analyze medical data, and with the help of human doctors can spot malignant tumours.
iPN: Cognitive Intelligence’s Application in Legal, Construction, Career Planning, and Architecture
iPIN focuses on cognitive intelligence that leverages socio-economic maps, and uses knowledge maps, semantic analysis, and deep learning technology to help machines understand words, syntax, interrelations, different scenarios, case studies, and multi-dimension rankings to grasp the human process of reasoning and analysis. iPIN is currently recruiting in the fields of HR, law, career planning, architecture and so forth. Yang Yang, iPIN Founder and CEO, will host a panel titled “Cognitive Analytics — Re-examine the Essence of Business Through Machines”
Cardinal Operations: In the era of AI, what can Operations Research do?
Cardinal Operations positions itself as a data science and data decision-making company. Through the description and observation of multi-dimensional data, data collection and management, deep learning optimization, data prediction, and mathematical models, Cardinal uses efficient optimization algorithms to help companies improve results. Services are applied to revenue, risk, supply chain management and so on.
For more information and an event schedule visit gmis.jiqizhixin.com