Kaisheng Ma
Tenured associate professor
Works on AI chiplets and accelerators, AI algorithms for self-driving, and implanted systems and neural dust.
Kaisheng Ma's lab at Tsinghua works on accelerators, chips & architecture and autonomous vehicles & drones. 47 papers since 2023, 32 at top venues, mostly AAAI and ICLR. 9 PhD students and 6 master's students. Applications close .
At a glance
- papers, 2023–2026
- 12/ year
- at T1 venues
- 68%
- PhD students
- 9
- MS students
- 6
- alumni listed
- 0
Papers by yearfrom the lab homepage, 2023–2026
- 5 papers2026 · T1 3 · collecting
- 8 papers2025 · T1 7
- 16 papers2024 · T1 11
- 18 papers2023 · T1 11
T1 venuesOtherVenues: AAAI, ICLR, BMVC, DAC, CVPR
Research focuspaper titles that use each topic's words
- Accelerators, chips & architecture7 papers
- Autonomous vehicles & drones0 papers
- self-driving0 papers
- interdisciplinary information0 papers
- iiis0 papers
Paper list
From the lab homepage.
20265 papers
- IEEE International Symposium on High-Performance Computer Architecture (HPCA), Industry TrackCharacterizing Cloud-Native LLM Inference at ByteDance and Exposing Optimization Challenges and Opportunities for Future AI Accelerators
- ICLR (T1)ES-dLLM: Efficient Inference for Diffusion Large Language Models by Early-Skipping
- DAC (T1)FBS: Accelerating CNN Inference over RNS-CKKS with Fewer Bootstrapping Sparsity
- the EAGE Annual Conference & Exhibition, oralPosition-based point-spread-function representation network for migration image deconvolution with regularised inversion
- DAC (T1)Throughput-Oriented Speculative Decoding via Intra-Device Parallelism
20258 papers
- IROS (T1)Benchmarking Long-Horizon Mobile Manipulation in Multi-Room Dynamic Environments
- DAC (T1)BufferProspector: Discovering and Exploiting Untapped Buffer Resources in Many-Core DNN Accelerators
- ICASSP (T1)Enhancing Vision: Harmonizing Frequency for Imaging Quality and Perception Accuracy
- Scientific ReportsHuman visual attention-inspired knowledge distillation underlying interpretable computational pathology
- ICML (T1)MindCustomer: Multi-Context Image Generation Blended with Brain Signal
- AAAI (T1)MindPainter: Efficient Brain-Conditioned Painting of Natural Images via Cross-Modal Self-Supervised Learning
- HPCA (T1)SoMa: ldentifying, Exploring, andUnderstanding the DRAMCommunication Scheduling Space forDNN Accelerators
- CCS (T1)WPC: Weight Plaintext Compression for CNN Inference based on RNS-CKKS
202416 papers
- ASPLOS (T1)Cocco: Hardware-Mapping Co-Exploration towards Memory Capacity Communication Optimization.
- AAAI (T1)CutFreq: Cut-and-Swap Frequency Components for Low-Level Vision Augmentation
- BMVCDRAFT: Direct Radiance Fields Editing with Composable Operations
- ICLR (T1)DreamLLM: Synergistic Multimodal Comprehension and Creation
- ATCEvaluating Chiplet-based Large-Scale Interconnection Networks via Cycle-Accurate Packet-Parallel Simulation
- AAAI (T1)Guiding a Harsh-Environments Robust Detector via RAW Data Characteristic Mining
- ICCAD (T1)LACO: A Latency-Constraint Offline Neural NetworkScheduler towards Reliable Self-DrivingPerception
- IEEE International Symposium on High-Performance Computer Architecture (HPCA) [Distinguished Artifact Award (1/410)]Mapping and Architecture Co-exploration for Large-scale DNN Chiplet Accelerators
- IROS (T1)MG-VLN: Benchmarking Multi-Goal and Long-Horizon Vision-Language Navigation with Language Enhanced Memory Map
- ICPRNAOL: NeRF-Assisted Omnidirectional Localization
- CVPR (T1)Orchestrate Latent Expertise: Advancing Online Continual Learning with Multi-Level Supervision and Reverse Self-Distillation
- ICCAD (T1)Partial Differential Equation Acceleration by Exploiting Value Similarity
202318 papers
- ISSCCA 28nm 68MOPS 0.18μJ/Op Paillier Homomorphic Encryption Processor with Bit-Serial Sparse Ciphertext Computing
- HPCA (T1)A Scalable Methodology for Designing Efficient Interconnection Network of Chiplets
- HCSA Scalable Multi‑Chiplet Deep Learning Accelerator with Hub‑Side 2.5D Heterogeneous Integration
- ICLR (T1)Autoencoders as Cross-Modal Teachers: Can Pretrained 2D Image Transformers Help 3D Representation Learning?
- ICML (T1)Contrast with Reconstruct: Contrastive 3D Representation Learning Guided by Generative Pretraining
- ICASSP (T1)CORSD: Class-Oriented Relational Self Distillation
- ICLR (T1)Hebbian and Gradient-based Plasticity Enables Robust Memory and Rapid Learning in RNNs
- MICRO (T1)Heterogeneous Die-to-Die Interfaces: Enabling More Flexible Chiplet Interconnection Systems
- ISCA (T1)Inter-layer Scheduling Space Definition and Exploration for Tiled Accelerators
- AAAI (T1)Language-Assisted 3D Feature Learning for Semantic Scene Understanding
- HCSPHEP: Paillier Homomorphic Encryption Processors for Privacy‑Preserving Applications in Cloud Computing
- CVPR (T1)PointDistiller: Structured Knowledge Distillation towards Efficient and Compact 3D Detection
Where alumni went
Alumni are not listed on the lab homepage.
Apply to TsinghuaFall 2027
Computer Science PhD (international applicants)
- Round
- Main
- Opens
- GRE
- See program page
- English
- Per school catalog
- Fee
- RMB 800
Not yet confirmed for this intake.
Similar labsnearest by research summary
- CNPeking
Cong Wang
- Recommendation, search & web
- CNSJTU
Shiyou Qian
- Data management & big data
- SGNTU
Gao Cong
- Data management & big data
- USUW
Amy Orsborn
- Neuroscience & brain-computer interfaces
- USGeorgia Tech
B. Aditya Prakash
- Data management & big data
- CNSJTU
Guanghui He
- Accelerators, chips & architecture
Your notes
Save this lab to keep notes.
Data
Checked against the lab homepage.
- Collected
- Verified
- Papers
- The lab homepage
- Members
- Lab homepage
- Alumni
- Lab homepage
Tiers are venue classifications, not lab rankings.
Tsinghuacloses