Bingsheng He
Professor
Builds data systems and infrastructure for large language models and agentic AI, federated learning and graph analytics on GPUs, TPUs and FPGAs.
Bingsheng He's lab at NUS works on accelerators, chips & architecture, distributed, cloud & edge systems and large language models. 134 papers since 2023, 110 at top venues, mostly SIGMOD and VLDB. 7 PhD students. 12 alumni, most often to Alibaba and A Startup Company in USA. Applications close .
At a glance
- papers, 2023–2026
- 35/ year
- at T1 venues
- 82%
- PhD students
- 7
- MS students
- 0
- alumni listed
- 12
Papers by yearfrom DBLP, 2023–2026
- 38 papers2026 · T1 32 · collecting
- 36 papers2025 · T1 29
- 33 papers2024 · T1 29
- 27 papers2023 · T1 20
T1 venuesOtherVenues: SIGMOD, VLDB, TKDE, ICDE, ICLR
Research focuspaper titles that use each topic's words
- Large language models25 papers
- Distributed, cloud & edge systems11 papers
- Accelerators, chips & architecture6 papers
- gpus4 papers
- graph analytics0 papers
Paper list
From DBLP.
202638 papers
- WWW (T1)'They've Stolen My GPL-Licensed Model!': Toward Standardized and Transparent Model Licensing (opens in a new tab)
- Frontiers Comput. Sci.A comprehensive taxonomy of prompt engineering techniques for large language models (opens in a new tab)
- SIGMOD (T1)ABFlow: Alert Bursting Flow Query in Streaming Temporal Flow Networks (opens in a new tab)
- ISCA (T1)Approaching Shannon Bound with Lossless LLM Weight Compression (opens in a new tab)
- EuroSys (T1)Arena: Efficiently Training Large Models via Dynamic Scheduling and Adaptive Parallelism Co-Design (opens in a new tab)
- SIGMOD (T1)Automating Database-Native Function Code Synthesis with LLMs (opens in a new tab)
- ASPLOS (T1)Bat: Efficient Generative Recommender Serving with Bipartite Attention (opens in a new tab)
- ICLR (T1)Beyond Prompt-Induced Lies: Investigating LLM Deception on Benign Prompts (opens in a new tab)
- TKDE (T1)BLITZ: A Flexible and Efficient Graph Mining System (opens in a new tab)
- IEEE Trans. Parallel Distributed Syst.ComStar: Compression-Aware Stream Query for Heterogeneous Hybrid Architecture (opens in a new tab)
- VLDB (T1)DBCooker: A Database Kernel-Specialized Coding Agent Harness (opens in a new tab)
- AAAI (T1)DGP: A Dual-Granularity Prompting Framework for Fraud Detection with Graph-Enhanced LLMs (opens in a new tab)
202536 papers
- TKDE (T1)A Survey of Imbalanced Learning on Graphs: Problems, Techniques, and Future Directions (opens in a new tab)
- ICDE (T1)Accelerating D-Core Maintenance over Dynamic Directed Graphs (opens in a new tab)
- HPCA (T1)AccelES: Accelerating Top-K SpMV for Embedding Similarity via Low-bit Pruning (opens in a new tab)
- ICML (T1)Adapting Precomputed Features for Efficient Graph Condensation (opens in a new tab)
- WWW (T1)Aggregate to Adapt: Node-Centric Aggregation for Multi-Source-Free Graph Domain Adaptation (opens in a new tab)
- ACM Comput. Surv.AI-powered Fraud Detection in Decentralized Finance: A Project Life Cycle Perspective (opens in a new tab)
- ICDE (T1)An Efficient Memoization Engine for Concurrent Graph Query Processing (opens in a new tab)
- KDD (T1)Blurred Encoding for Trajectory Representation Learning (opens in a new tab)
- IEEE Trans. Cloud Comput.Breaking the Edge: Enabling Efficient Neural Network Inference on Integrated Edge Devices (opens in a new tab)
- SIGMOD (T1)Bursting Flow Query on Large Temporal Flow Networks (opens in a new tab)
- SIGMOD (T1)Clementi: Efficient Load Balancing and Communication Overlap for Multi-FPGA Graph Processing (opens in a new tab)
- SIGMOD (T1)Community Detection in Heterogeneous Information Networks Without Materialization (opens in a new tab)
202433 papers
- TKDE (T1)A Survey on Concurrent Processing of Graph Analytical Queries: Systems and Algorithms (opens in a new tab)
- VLDB WorkshopBandwidth Expansion via CXL: A Pathway to Accelerating In-Memory Analytical Processing (opens in a new tab)
- WWW (T1)Collaborate to Adapt: Source-Free Graph Domain Adaptation via Bi-directional Adaptation (opens in a new tab)
- ICLR (T1)Consistency Training with Learnable Data Augmentation for Graph Anomaly Detection with Limited Supervision (opens in a new tab)
- EMNLP (T1)CryptoTrade: A Reflective LLM-based Agent to Guide Zero-shot Cryptocurrency Trading (opens in a new tab)
- ICLR (T1)Effective and Efficient Federated Tree Learning on Hybrid Data (opens in a new tab)
- TKDE (T1)Efficient Heterogeneous Graph Learning via Random Projection (opens in a new tab)
- SIGMOD (T1)Enabling Adaptive Sampling for Intra-Window Join: Simultaneously Optimizing Quantity and Quality (opens in a new tab)
- ICLR (T1)EX-Graph: A Pioneering Dataset Bridging Ethereum and X (opens in a new tab)
- AAAI (T1)Exploiting Label Skews in Federated Learning with Model Concatenation (opens in a new tab)
- ASPLOS (T1)FaaSGraph: Enabling Scalable, Efficient, and Cost-Effective Graph Processing with Serverless Computing (opens in a new tab)
- NeurIPS (T1)Federated Transformer: Multi-Party Vertical Federated Learning on Practical Fuzzily Linked Data (opens in a new tab)
202327 papers
- VLDB (T1)A Design Space Exploration and Evaluation for Main-Memory Hash Joins in Storage Class Memory (opens in a new tab)
- ICDE (T1)A High-Performance Index for Real-Time Matrix Retrieval (Extended Abstract) (opens in a new tab)
- Math. Comput.A rank-two relaxed parallel splitting version of the augmented Lagrangian method with step size in (0,2) for separable convex programming (opens in a new tab)
- TKDE (T1)A Survey on Federated Learning Systems: Vision, Hype and Reality for Data Privacy and Protection (opens in a new tab)
- ICDE (T1)A Survey on Spark Ecosystem: Big Data Processing Infrastructure, Machine Learning, and Applications (Extended abstract) (opens in a new tab)
- ICML (T1)Adversarial Collaborative Learning on Non-IID Features (opens in a new tab)
- WWW (T1)BERT4ETH: A Pre-trained Transformer for Ethereum Fraud Detection (opens in a new tab)
- ICPPCommunication-Efficient Generalized Neuron Matching for Federated Learning (opens in a new tab)
- SIGMOD (T1)DeltaBoost: Gradient Boosting Decision Trees with Efficient Machine Unlearning (opens in a new tab)
- ICDE (T1)EdgeNN: Efficient Neural Network Inference for CPU-GPU Integrated Edge Devices (opens in a new tab)
- TKDE (T1)Efficient Decomposition Selection for Multi-class Classification (opens in a new tab)
- VLDB (T1)FEBench: A Benchmark for Real-Time Relational Data Feature Extraction (opens in a new tab)
Where alumni went12 listed, 70% to industry
Industry 7Academia 3
Most often to
- Alibaba
- A Startup Company in USA
- Huawei
Apply to NUSFall 2027
School of Computing PhD in Computer Science
- Round
- Main
- Opens
- —
- GRE
- See program page
- English
- See program page
- Fee
- See program page
Not yet confirmed for this intake.
Similar labsnearest by research summary
- CNZhejiang
Xiaohang Wang
- Accelerators, chips & architecture
- CNSJTU
Zeyu Mi
- Efficient training & inference
- CHETH Zurich
Shivaram Venkataraman
- Distributed, cloud & edge systems
- USUCSD
Yiying Zhang
- Efficient training & inference
- HKCUHK
Evangeline F. Y. Young
- Accelerators, chips & architecture
- USUW
Baris Kasikci
- Efficient training & inference
Your notes
Save this lab to keep notes.
Data
Checked against the lab homepage.
- Collected
- Verified
- Papers
- DBLP
- Members
- Lab homepage
- Alumni
- Lab homepage
Tiers are venue classifications, not lab rankings.
NUScloses