Haotian Tang

Haotian Tang

Auto-generated profile, based on publicly available information

Haotian Tang

Massachusetts Institute of Technology (MIT), 77 Massachusetts Avenue, Cambridge, MA 02139, USA

Research Focus & Metrics

Academic Career Length

5Years (2022-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
28 October 2023
Verified
Conference Paper
Open access
HTHaotian Tang
Haotian Tang
SYShang Yang
Shang Yang
ZLZhijian Liu
Zhijian Liu
KHKe Hong
Ke Hong
ZYZhongming Yu
Zhongming Yu
SHSong Han
Song Han
Sparse convolution plays a pivotal role in emerging workloads, including point cloud processing in AR/VR, autonomous driving, and graph understanding in recommendation systems. Since the computation pattern is sparse and irregular, specialized high-performance kernels are required. Existing GPU libraries offer two dataflow types for sparse convolution. The gather-GEMM-scatter dataflow is easy to implement but not optimal in performance, while the dataflows with overlapped computation and memory access (e.g. implicit GEMM) are highly performant but have very high engineering costs. In this paper, we introduce TorchSparse++, a new GPU library that achieves the best of both worlds. We create a highly efficient Sparse Kernel Generator that generates performant sparse convolution kernels at less than one-tenth of the engineering cost of the current state-of-the-art system. On top of this, we design the Sparse Autotuner, which extends the design space of existing sparse convolution libraries and searches for the best dataflow configurations for training and inference workloads. Consequently, TorchSparse++ achieves 2.9 × , 3.3 × , 2.2 × and 1.7 × measured end-to-end speedup on an NVIDIA A100 GPU over state-of-the-art MinkowskiEngine, SpConv 1.2, TorchSparse and SpConv v2 in inference; and is 1.2-1.3 × faster than SpConv v2 in mixed precision training across seven representative autonomous driving benchmarks. It also seamlessly supports graph convolutions, achieving 2.6-7.6 × faster inference speed compared with state-of-the-art graph deep learning libraries. Our code is publicly released at https://github.com/mit-han-lab/torchsparse.

6 September 2022 in Zhonghua Yi Xue Za Zhi
Verified
Journal Article
Closed access
XWXiangbin Wang
Xiangbin Wang
HTH Tang
H Tang
YCYa-Jung Cheng
Ya-Jung Cheng
HSH B Shang
H B Shang
JMJ G Ma
J G Ma
ZWZhebao Wu
Zhebao Wu
The current study aimed to investigate the clinical feasibility of microscopic resection of hemilateral tuberculum sellae meningiomas (TSM) via the contralateral eye brow arch approach. The clinical data of 34 patients with TSM who underwent microsurgery from January 2016 to June 2021 in the Neurosurgery Department of Ruijin Hospital Affiliated to Shanghai Jiaotong University School of Medicine and the First Affiliated Hospital of Henan University were collected and reviewed. The postoperative visual acuity improvement rate was 88.5% (23/26), and the total tumor resection rate was 88.2% (30/34); the postoperative visual acuity improvement in patients with total tumor resection was better than that of patients with partial resection [90.9% (20/22) vs 3/4]. Meanwhile, the postoperative visual acuity improvement in patients with the superior optic nerve and laterl-superior optic nerve was better than that of patients with the lateral optic nerve type (12/14, 8/8 vs 3/4). Supraorbital skin numbness occurred in 3 cases after operation, and the symptoms disappeared during follow-up; 2 cases had mild disturbance of hormone level, and urine output of 2 cases increased after operation, which returned to normal level after symptomatic treatment; 1 case had subcutaneous effusion which was absorbed after treatment. There were no complications such as olfactory disturbance and intracranial infection. During follow-up for 3-60 (33±6) months, recurrence occurred in 2 cases and reoperation was performed. For the hemilateral TSM, according to the preoperative evaluation of the origin of the TSM and the side with visual impairment, the contralateral eyebrow approach is selected to fully expose the tumor base below the optic nerve. It is beneficial to fully resect the tumor under direct vision, and the symptoms of postoperative visual impairment are significantly improved, indicating that the current surgical method can be used in the clinical setting.

1 January 2024 in World Scientific Annual Review of Artificial Intelligence
Verified
Journal Article
Closed access
ZLZhijian Liu
Zhijian Liu
HTHaotian Tang
Haotian Tang
YLYujun Lin
Yujun Lin
SHSong Han
Song Han

21 October 2022 in arXiv
Verified
Preprint
Open access
PDF file
XWXueru Wen
Xueru Wen
CZChangjiang Zhou
Changjiang Zhou
HTHaotian Tang
Haotian Tang
LLLuguang Liang
Luguang Liang
YJYu Jiang
Yu Jiang
HQHong Qi
Hong Qi
Named entity recognition is a fundamental task in natural language processing, identifying the span and category of entities in unstructured texts. The traditional sequence labeling methodology ignores the nested entities, i.e. entities included in other entity mentions. Many approaches attempt to address this scenario, most of which rely on complex structures or have high computation complexity. The representation learning of the heterogeneous star graph containing text nodes and type nodes is investigated in this paper. In addition, we revise the graph attention mechanism into a hybrid form to address its unreasonableness in specific topologies. The model performs the type-supervised sequence labeling after updating nodes in the graph. The annotation scheme is an extension of the single-layer sequence labeling and is able to cope with the vast majority of nested entities. Extensive experiments on public NER datasets reveal the effectiveness of our model in extracting both flat and nested entities. The method achieved state-of-the-art performance on both flat and nested datasets. The significant improvement in accuracy reflects the superiority of the multi-layer labeling strategy.