- 목적
마이크로소프트연구소아시아와 공동연구협력 프로젝트* 선정
* 2020년 글로벌 핵심인재 양성지원 사업(과기정통부)의 ‘글로벌 기업 연계형’과제로 추진되며, 본 프로젝트는 ‘국가연구개발사업의 관리 등에 관한 규정 제32조의 2’에 의거하여 동시에 수행 가능한 최대 과제 수에 해당함.
- 운영방향
마이크로소프트연구소아시아에서 선정한 연구주제에 부합하는 창의적 아이디어 공모를 통해 선정
프로젝트 수행시 공동연구협력을 위한 마이크로소프트연구소아시아의 전문가 매칭
공동 연구원* 마이크로소프트연구소아시아에 6개월 파견
* 공동연구기관(국내 대학)의 연구원(석박사생)
- 과정개요
A. 프로젝트 지원
– 지원규모: 총 12-13억 원*, 10-12개 과제
* 글로벌 핵심인재 양성지원 사업(과기정통부) 예산 : 10억 원
– 지원분야:
Content Retrieval and Generation for Searching, Reading, Writing, Coding, Watching and Listening
Retrieving or generating accurate, attractive and timely contents is crucial to many real-life scenarios, such as news reports (e.g., Toutiao and MS News), search engine (e.g., Google and Bing) and content editor (e.g., Office and Visual Studio). Motivated by this, we propose content retrieval and generation as a research theme and call for collaborations with professors/researchers from academic community. The goal is to develop cutting-edge models for various tasks, such as title generation for news and ads, code retrieval and generation, text summarization and enrichment, cross-modal content retrieval, speech-to-speech translation for audio contents, question answering based on heterogeneous contents, etc. From research perspective, we hope the collaborations can lead to impactful research papers or achieve state-of-the-art results on latest research-driven leaderboards.
Topics:
- Text Generation for Search, News and Ads
- Commonsense Extraction, Reasoning, QA
- Image Retrieval, QA and Captioning
- Code Retrieval and Generation
- Speech-to-Speech Translation
- Video Retrieval, QA and Captioning
Next Generation Recommendation Systems
Information overload has become a huge challenge for online users. In order to alleviate this problem, recommendation systems play an increasingly important role in Internet services. It is a constant hot topic in industry and academia. In this theme, we encourage researchers to collaborate and tackle important challenges in recommendation systems, including data heterogeneity, data sparsity, and lack of interpretability. We welcome research proposals which aim to improve the recommendation performance by leveraging recent progress in deep learning, natural language understanding, and knowledge graph. Priority will be given to proposals which could contribute to Microsoft Recommenders , an open source repository for helping developers to build their own recommendation systems more efficiently. We believe that personalized recommendation systems will continue to develop in various directions, including effectiveness, diversity, computational efficiency, and interpretability, ultimately addressing the problem of information overload. Please refer to our recent article for more details about our vision.
Topics:
- Deep learning based recommendation
- Deep learning based user representation
- Explainable Recommendation Systems
- Reinforcement learning based recommendation
- Knowledge aware recommendation
- Transfer learning based recommendation
Advanced Machine Learning algorithms
Thanks to big labeled data, big model, and big computing power, deep learning and reinforcement learning have achieved great success in recent years. However, big data/model/computation also bring several key challenges for machine learning to move to next stage. To address those challenges and further advance machine learning research, We welcome research proposals which aim to develop low-resource (learning from less labeled data), fast and/or lightweight algorithms for speech, natural language, or computer vision. Priority will be given to proposals which could contribute to Microsoft Dual Learning or Microsoft MASS. We believe that low-resource, fast and lightweight learning algorithms will enable the applications of latest machine learning techniques in various industries and scenarios.
Topics:
- Speech generation/transfer/translation
- Transfer learning
- Multi-agent reinforcement learning
- Reinforcement learning with demonstration
3D Content Creation for Learning and from Learning
As deep learning demonstrates its capability for 2D image/video generation, deep learning based 3D content creation still faces several challenges. One challenge is the limited amount of 3D data in the existing dataset, which limits research and development of deep learning algorithm in this field. Another challenge is how to fuse deep learning-based 3D content creation methods with existing interactive modeling approaches for variant 3D modeling/animation tasks. In this research theme, we aim to develop a set of dataset and labeling tools, as well as algorithms to tackle these two challenges.
Topics:
- High quality 3D object database acquisition
- Data labeling tools and system
- Learning based 3D content creation algorithms/network design
Interpretable, Transparent and Safe Deep Learning for Computer Vision
Deep learning has brought tremendous progress in computer vision applications, such as visual recognition, image translation, and multi-modal analysis. However, in contrast to traditional machine learning models such as AdaBoost and support vector machines, deep learning models are often perceived as black boxes that offer few clues as to how they reach their results, potentially raising safety concerns. For this theme, we solicit research proposals on interpretable, transparent and safe deep learning for computer vision, including topics such as but not limited to interpreting deep learning models, designing transparent deep learning models, explicitly embedding human knowledge into deep learning models, and building safe and reliable deep learning models. Proposals that involve collaboration with MSR Asia computer vision researchers are highly encouraged. Priority will be given to proposals with outputs, e.g., research papers, demos, or open source repositories, aligned with Microsoft AI principles.
Topics:
- Interpret deep learning models
- Adversarial examples
- Interpretable neural architecture design
- Transparent deep learning for content creation
Human Understanding in Video
We are developing both cloud-based and edge-based intelligence engines that can turn raw video data into insights to facilitate various applications and services such as business (retail store, office) intelligence, smart home intelligence, video augmented reality, etc. We have taken a human centric approach that focuses on understanding human, human attributes, and human activities in the scene. Successful applications are required to commit to open source release of their outcomes as part of a collective Human SDK open source project.
Topics:
- Human detection and tracking
- Human re-ID
- Human pose estimation
- Human action recognition
- 2D/3D scene understanding
- Multi-modality (audio/visual/language) human understanding
Large-Scale Networking Platform for Machine Learning Based Real Time Communications
Leveraging AI to boost real-time communication (RTC) application has attracted much attention in academia and industry. However, training AI models for RTC application demands massive amounts of real Internet data. Until now, these data are far from enough to train an ideal model for RTC application since RTC is mainly quality of experience (QoE)-driven and requires precise network estimations. MSRA is calling for an open global RTC research platform involving many universities and research institutions. Such platform, which we call it OpenNetLab, provides an RTC evaluation service with network testbed, composed of heterogeneous endpoints all around the world. OpenNetLab will be implemented by virtualization technologies, such as VMs and containers. Researchers in the community can submit their application evaluation jobs or network measurement jobs by themselves. By building and operating this project, universities can cooperate the following topics with Microsoft Research Asia and other universities, including but not limited to: Building network stat collection platform; Realistic Network Model; Machine Learning-based RTC Protocol; Light-weight bandwidth estimation technique, etc. Welcome to participate in OpenNetLab project to build nodes and conduct research as well as evaluate applications in their campus.
Topics:
- Global Network Stat Collection Platform
- Realistic Network Model
- Networking Gym for RTC applications
- ML/RL algorithm and (h/w-based)
- acceleration for RTC Protocol
- Lightweight bandwidth estimation technique
Infuse AI to Empower Heterogeneous Devices and New Applications on the Edge
With the recent advances in software and hardware, there is computing paradigm shift from centralized cloud computing to distributed computing on the edge. Together with the breakthroughs in AI, we advocate intelligent edge computing to infuse AI to empower heterogeneous devices and diverse AI applications on the edge. We call for proposals on intelligent edge computing that 1) advance the state-of-the-art research by top publications, prototypes and open-source code; 2) build and deploy real edge systems to solve real problems and learn experience in the wild; and 3) leverage Azure cloud to connect and enable new generation devices and applications.
Topics:
- Affordable AI models tailored for diverse hardware (model compression and optimization, AutoML etc.)
- Efficient software stack best utilizing heterogenous resources (system optimization, resource management and scheduling, software/hardware co-design etc.)
- Privacy and security (user privacy, data and model protection etc.)
- Learning on the edge (distributed learning, continuous learning, collaborative learning etc.)
- New applications and scenarios (AIoT, AR, VR, gaming, 5G etc.)
– 지원기간: 2020.3.25 – 2021.5.31
– 지원내용: 프로젝트경비 (기업부문: USD20K와 정부부문: Korea Won80-100M) – 총 1억 – 1억 3천만원
* 정부부문의 프로젝트비 산정. 사용 등은 정보통신방송 연구개발 관리규정에 따름 (IITP의 추후 안내)
* 기업부문은 기업과제 별도 계약에 따름.
– 선정심사: 마이크로소프트연구소아시아의 전문성심사(서면)
* 심사결과는 개별 통보하며 공개되지 않음
* 과제 공동연구기관으로 선정된 국내 대학은 IITP와 협약 등 정부 과제 수행을 위한 절차를 추진해야 함
B. 공동연구원 파견
– 공동 연구원: 별도 심사를 통해 선발
– 파견기간: 6개월(2020. 9월 – 2021. 3월 예정)
– 파견기관: 마이크로소프트연구소아시아 (중국, 북경)
- 신청자격
프로젝트당 학생(2-5명) 및 지도교수로 팀을 구성
학생: 국내 IT관련학과 대학원에 재학중인 전일제 석박사과정 대학원생
교수: 국내 IT관련학과 소속 전임교원으로서 지원기간 동안 프로젝트 총괄 및 학생 연구 지도가
가능한 자
- 지원절차
프로젝트 선정 공고 -> 제안서 제출(온라인 접수, 지원양식, 100% 영문제안) -> 선정심사 -> 지원대상 선정통보 -> 협약체결 및 프로젝트 경비 지급
- 신청 유의사항
프로젝트팀은 총1개 분야에 한해 신청할 수 있음
신청자격에 부합하지 않을 경우 선정심사 대상에서 제외될 수 있음
* 국가연구개발사업에 참여제한 중인 자와 기관은 신청 불가
- 신청요령
신청방법: 이메일 신청([email protected])
신청접수마감: 2020년 3월 13일(금) 17:00
* 제출된 서류는 일제 반환되지 않음
- 문의처
사업담당: 마이크로소프트연구소 이미란 전무 (010-3600-4227, [email protected])
