>

Research Areas

Computer Vision

Focusing through machine learning enables machines to have the ability to sense and understand the world around them. Specific research directions include: visual image enhancement, semantic understanding and reasoning of images.

Search & Recommendation

Noah’s Ark Lab has developed recommendation systems and search engine with the most advanced technologies in the industry. The recommendation system and search engine, equipped with large scale machine learning and NLP technologies, can provide high recommendation and search accuracy in any given environment.

AI Theory

We focus on fundamental learning algorithms and theories, but we also explore possible industrial applications. We are interested in AI systems that can learn and reason from small data, adapt over time, collaborate with each other, and interact with humans. We are actively looking for talents who are passionate about fundamental research and real-world applications.

Speech and Language Processing

Enriching people’s lives through communication is Huawei’s vision for the future. Technologies for speech and language processing are fundamental and important to achieve this goal. We are developing advanced technologies of speech and language processing at Noah’s Ark Lab. Research topics include: speech recognition and synthesis, machine translation, natural language dialogue, natural language generation and question answering. We have been extensively working on building deep learning systems that can communicate with humans and help them on real-world tasks, for examples, neural machine translation and neural dialogue systems. We wish to jointly work with the community to advance the state of the art in the fields as well as to harness the products and services of the company.

Decision Making & Reasoning

Ranging from anomaly prediction, through operations research, to deep reinforcement learning, our work is about making good decisions and learning to make better decisions. We support Huawei’s business by making our global technical service (GTS) and supply chain management orders-of-magnitude more efficient, by understanding for a smoothly connected world, and by building the smarts of safe and connected cars for roads everywhere.

Research Projects

Terminal Vision

Focusing on AI algorithm design for new HUAWEI terminals visual experience, including intelligent scene recognition, saliency detection, image semantic segmentation, image color balance, image super resolution and denoising, OCR, image search, image captioning and so on.

Net Mind

Focus on the application of AI technology in telecommunication networks. Using key technologies such as reinforcement learning, deep learning, robust AI, and spatial-temporal prediction, to implement network space-time prediction and optimization control, and finally to achieve an automatic, self-healing, and self-optimized automatic driving network.

Enterprise Intelligence

Focusing on data analysis and management, we leverage AI algorithms, such as demand prediction, inventory control and multi-object optimization, to support and optimize daily operation at enterprise level. Solutions include but not limit to smart supply chain, smart finance system and smart HR system.

Machine Learning for Recommendation & Search

We focus on machine learning driven recommendation and search technologies, the research directions includes deep learning for recommendation system, reinforcement learning for search, online learning technology, Exploration & Exploitation, search query recommendation, new word detection, the general low rank models, such as one class matrix factorization, field-aware factorization machine

Safe City

Based on crowd-aware monitoring video content analysis tasks, we focus on challenging issues such as “unclear”, “not sure”, “unable to found”, and “number wrong” in video surveillance, by breaking through the key technologies of image quality enhancement, human attribute recognition, monitoring face recognition, multimode biometric fusion, video event analysis, etc.. The solutions will be verified in real-world business scenarios and iterated to promote technology commercially.

Huawei Voice Assistant

Communicating with smart devices through voice and natural language has become one of the major way of human-device interaction. At Noah’s Ark lab, we have been working with the teams at Huawei consumer business group to build the intelligent voice assistant on Huawei mobile devices. We are developing a series of technologies such as far-field speech recognition, task-oriented multi-turn dialogue systems, neural machine translation, to enhance the interaction experiences of Huawei’s smartphones. Some technologies have been shipped to Huawei smartphones since the Mate 10 series release. In the future, we will continue to develop more new technologies and apply them to more Huawei’s smart devices.

GTS Brain

Focus on artificial intelligence algorithms and systems, promote intelligent and smart transformation of network O&M, and lay a solid foundation for future O&M services of tens of billions of dollars in inventory networks. Specifically includes an industrial AR assistant system constructed based on machine vision such as target detection, SLAM, and target tracking, to improve network operation efficiency. Based on the modeling and analysis of all telecom fields and fine-grained data, accurate prediction of network spatiotemporal traffic is implemented, and intelligent network management, planning, and upgrade are implemented. Based on sequence data prediction, anomaly detection, and root cause analysis, intelligent fault locating and quick recovery of network problems are implemented. Finally, Precision, automation, and robustness of the network service quality assurance system are achieved.