Science Explorer Interactive view Map
Company · Beijing · China

Microsoft Research Asia (China)

In research, Microsoft Research Asia (China) stands highest in Computer Science (#77 of 4,667 worldwide), Artificial Intelligence (#131 of 1,884 worldwide) and Computer Vision and Pattern Recognition (#72 of 941 worldwide), 2022–2025. Relative to its size it is most specialised in Computer Vision and Pattern Recognition, Signal Processing and Artificial Intelligence — Computer Vision and Pattern Recognition is 19.4× its share of world research.

research.microsoft.com/en-us/labs/asiaROR OpenAlex
World rank, 2022–2025
#2,078
of 28,054 · #1,409 all time
Rank in China
#554
of 3,058
Research works
3.6k
▲ 12% vs 2013–17
Citations
271k
75.9 per fractional work
Top-10% rate
39.1%
record average 16.5%
Open access
20%
world 28%

What is Microsoft Research Asia (China) known for in research?

The fields where it stands highest, 2022–2025, ranked among every institution above the floor in each field.

FieldWorld rankWhere that sitsTop-10% rateWorksAll time
Computer ScienceField #77 of 4,667 38.4%434 #64
Artificial IntelligenceSubfield #131 of 1,884 34.2%182 #96
Computer Vision and Pattern RecognitionSubfield #72 of 941 45.7%134 #36
Physical SciencesDomain #1,276 of 12,888 39.0%497 #1008

Each strip is that field’s whole ranked pool, with the notch where Microsoft Research Asia (China) sits in it, in the colour of the band that rank falls in. The track under the rate is the rate itself: it carries no world mark, because the world rate differs by field (from about 6% to 21% in this record).

Which of those standings moved

The same fields on both windows: where it stood over the whole record, and where it stands in 2022–2025. The distance is the story, and it is the one thing the two rank columns above cannot show.

  1. Computer Science#64 #77
  2. Artificial Intelligence#96 #131
  3. Computer Vision and Pattern Recognition#36 #72
  4. Physical Sciences#1,008 #1,276
all time2022–2025, better2022–2025, worse

A dot further right is a better standing. Fields it was not ranked in over the whole record are left out.

What does Microsoft Research Asia (China) specialise in?

Where its research is concentrated relative to its size: Computer Vision and Pattern Recognition takes 19.4× the share of its output that it takes of world research.

  1. Computer Vision and Pattern RecognitionSubfield · 133.7 works19×
  2. Signal ProcessingSubfield · 32.1 works19×
  3. Artificial IntelligenceSubfield · 181.9 works14×
  4. Computer ScienceField · 434.3 works10×
  5. Computer Networks and CommunicationsSubfield · 27.3 works5.6×
← less than its size predicts1×more →

Location quotient, a volume reading rather than an impact one. It surfaces small, lopsided specialities.

Field profile

Location quotient across every field it publishes in: outside the ring is more than an institution of this size would be expected to publish, inside it is less.

Agricultural and Biological Sciences: 0.01× its world share, 0 worksAgricultural and Bio…Biochemistry, Genetics and Molecular Biology: 0.31× its world share, 9 worksBiochemistry, Geneti…Immunology and Microbiology: 0.01× its world share, 0 worksImmunology and Micro…Neuroscience: 0.39× its world share, 4 worksNeuroscienceChemical Engineering: 0.10× its world share, 0 worksChemical EngineeringChemistry: 0.08× its world share, 1 worksChemistryComputer Science: 10.23× its world share, 434 worksComputer ScienceEarth and Planetary Sciences: 0.37× its world share, 3 worksEarth and Planetary …Energy: 0.01× its world share, 0 worksEnergyEngineering: 0.51× its world share, 42 worksEngineeringEnvironmental Science: 0.15× its world share, 4 worksEnvironmental ScienceMaterials Science: 0.25× its world share, 5 worksMaterials ScienceMathematics: 0.15× its world share, 1 worksMathematicsPhysics and Astronomy: 0.56× its world share, 7 worksPhysics and AstronomyDentistry: 0.03× its world share, 0 worksDentistryHealth Professions: 0.05× its world share, 1 worksHealth ProfessionsMedicine: 0.07× its world share, 8 worksMedicineNursing: 0.02× its world share, 0 worksNursingArts and Humanities: 0.04× its world share, 1 worksArts and HumanitiesBusiness, Management and Accounting: 0.24× its world share, 4 worksBusiness, Management…Decision Sciences: 1.96× its world share, 10 worksDecision SciencesEconomics, Econometrics and Finance: 0.11× its world share, 1 worksEconomics, Econometr…Psychology: 0.19× its world share, 3 worksPsychologySocial Sciences: 0.11× its world share, 8 worksSocial Sciences1×world share
more than its size predictsabout as predictedless

Rings at 0.5×, 1× and 2×. Widest outward: Computer Science, 10.2×. Every wedge is a field page.

Who are the top researchers at Microsoft Research Asia (China)?

Ranked on the composite score, Tie‐Yan Liu, Yu Zheng and Xing Xie lead among researchers whose main affiliation is Microsoft Research Asia (China).

  1. 1 Tie‐Yan Liu China · #2,227 worldwide 3k citations · 148 works
  2. 2 Yu Zheng China · #2,339 worldwide 2.5k citations · 78 works
  3. 3 Xing Xie China · #2,863 worldwide 2.5k citations · 188 works
  4. 4 Wei Chen China · #3,368 worldwide 3.5k citations · 231 works
  5. 5 Wei‐Ying Ma China · #6,149 worldwide 1.9k citations · 100 works
  6. 6 Jingdong Wang China · #13,295 worldwide 2.3k citations · 171 works
  7. 7 Chin-Yew Lin China · #13,311 worldwide 784 citations · 58 works
  8. 8 Jizheng Xu China · #13,426 worldwide 940 citations · 70 works
  9. 9 Stephen Lin China · #13,554 worldwide 974 citations · 65 works
  10. 10 Ming Zhou China · #15,578 worldwide 1.3k citations · 90 works
All ranked researchers at Microsoft Research Asia (China)

Which keywords describe research at Microsoft Research Asia (China)?

By fractional works in 2022–2025, weighted toward what it does more of than the world: Deep Learning, Neural Networks, Convolutional Neural Networks, Information Retrieval, Text Classification, Topic Modeling, Machine Translation and Semantic Similarity.

▲ more of its work than of the world’s◆ about the world’s share▼ less than the world’s

Size is fractional works in 2022–2025 in the topics tagged with each word; colour is the word's share of this institution's work against its share of the world's. The 40 words are chosen for being large and distinctive. Each links to the topic it comes from most.

All 40 words, with their numbers
  1. Deep Learning207▲ 11×66 topics
  2. Neural Networks148▲ 18×28 topics
  3. Convolutional Neural Networks58▲ 14×13 topics
  4. Machine Learning50▲ 2.8×55 topics
  5. Information Retrieval45▲ 38×4 topics
  6. Text Classification42▲ 49×2 topics
  7. Topic Modeling40▲ 50×2 topics
  8. Machine Translation40▲ 44×2 topics
  9. Semantic Similarity40▲ 60×1 topic
  10. Word Representation40▲ 60×1 topic
  11. Unsupervised Learning38▲ 49×3 topics
  12. Representation Learning37▲ 48×3 topics
  13. Corpus Linguistics30▲ 26×2 topics
  14. Neural Machine Translation30▲ 57×1 topic
  15. Statistical Machine Translation30▲ 57×1 topic
  16. Semi-Supervised Learning29▲ 48×3 topics
  17. Convolutional Networks28▲ 36×3 topics
  18. Deep Neural Networks24▲ 40×3 topics
  19. Image Captioning22▲ 83×1 topic
  20. Visual Question Answering22▲ 51×1 topic
  21. Meta-Learning22▲ 49×2 topics
  22. Transfer Learning20▲ 19×3 topics
  23. Generative Adversarial Networks18▲ 40×2 topics
  24. Big Data17▲ 3.5×13 topics
  25. Domain Adaptation17▲ 63×1 topic
  26. Few-Shot Learning17▲ 63×1 topic
  27. Hidden Markov Models17▲ 53×2 topics
  28. Acoustic Modeling17▲ 81×1 topic
  29. Feature Extraction17▲ 6.6×9 topics
  30. Speaker Verification17▲ 81×1 topic
  31. Anomaly Detection16▲ 10×4 topics
  32. Distributed Systems16▲ 17×6 topics
  33. Image Classification14▲ 15×2 topics
  34. Collaborative Filtering13▲ 68×1 topic
  35. Matrix Factorization13▲ 68×1 topic
  36. Pose Estimation13▲ 21×2 topics
  37. Computer Vision12▲ 8.3×4 topics
  38. Beamforming12▲ 20×4 topics
  39. Depth Estimation12▲ 26×2 topics
  40. Image Recognition12▲ 16×1 topic

Which research topics does Microsoft Research Asia (China) publish most on?

By volume in 2022–2025: Topic Modeling, Natural Language Processing Techniques, Multimodal Machine Learning Applications and Domain Adaptation and Few-Shot Learning.

Topic Modeling40Topic Modeling: 40 fractional worksNatural Language Processing Techniques: 30 fractional worksMultimodal Machine Learning Applications: 22 fractional worksDomain Adaptation and Few-Shot Learning: 17 fractional worksSpeech Recognition and Synthesis: 17 fractional worksRecommender Systems and …13Recommender Systems and Techniques: 13 fractional worksAdvanced Neural Network Applications: 12 fractional worksSpeech and Audio Processing: 11 fractional worksAdvanced Vision and Imaging: 11 fractional worksMusic and Audio Processing: 10 fractional works
Artificial IntelligenceComputer Vision and Pattern RecognitionInformation SystemsSignal Processing

Area is fractional works; colour is the subfield each topic belongs to.

  1. 1 Topic Modeling Artificial Intelligence 40 works
  2. 2 Natural Language Processing Techniques Artificial Intelligence 30 works
  3. 3 Multimodal Machine Learning Applications Computer Vision and Pattern Recognition 22 works
  4. 4 Domain Adaptation and Few-Shot Learning Artificial Intelligence 17 works
  5. 5 Speech Recognition and Synthesis Artificial Intelligence 17 works
  6. 6 Recommender Systems and Techniques Information Systems 13 works
  7. 7 Advanced Neural Network Applications Computer Vision and Pattern Recognition 12 works
  8. 8 Speech and Audio Processing Signal Processing 11 works
  9. 9 Advanced Vision and Imaging Computer Vision and Pattern Recognition 11 works
  10. 10 Music and Audio Processing Signal Processing 10 works

How open and international is its research?

Against the world’s own shares — the tick on each track. Both are shares of its output, so they sit on one scale and can be read against each other as well as against the world.

openly available19.8%not open80.2%

World: 28% of research is openly available.

with international co-authors42.5%domestic only57.5%

World: 19% is written across borders.

How has Microsoft Research Asia (China)'s research output changed?

Output in 2018–2022 was 12% higher than in 2013–2017.

20002005201020152020
grewheldshrank

The same series as a ribbon — one cell per year, darker for more. The line above answers how much; this answers when.

Other research institutions in Beijing

All research institutions in Beijing

Research measures only: rankings here say nothing about teaching, admissions or student experience. Comparable institutions and collaboration partners are in the interactive view on the map.