How GPU Computing can Accelerate the Treatment of ... · PASCAL-VOC Test set Object detection on...
Transcript of How GPU Computing can Accelerate the Treatment of ... · PASCAL-VOC Test set Object detection on...
![Page 1: How GPU Computing can Accelerate the Treatment of ... · PASCAL-VOC Test set Object detection on PASCAL-VOC Test set Classification on ImageNet ‘val’ set Effect of pre-training](https://reader034.fdocuments.us/reader034/viewer/2022042606/5f7f4a35fc5c7a39192a1a8c/html5/thumbnails/1.jpg)
How GPU Computing can Accelerate the Treatment of Neurological Disorders
Eric K Oermann, MDAnthony B Costa, PhD
Icahn School of Medicine at Mount Sinai
![Page 2: How GPU Computing can Accelerate the Treatment of ... · PASCAL-VOC Test set Object detection on PASCAL-VOC Test set Classification on ImageNet ‘val’ set Effect of pre-training](https://reader034.fdocuments.us/reader034/viewer/2022042606/5f7f4a35fc5c7a39192a1a8c/html5/thumbnails/2.jpg)
Disclosures● EKO reports no relevant financial conflict of interest● ABC reports no relevant financial conflict of interest
![Page 3: How GPU Computing can Accelerate the Treatment of ... · PASCAL-VOC Test set Object detection on PASCAL-VOC Test set Classification on ImageNet ‘val’ set Effect of pre-training](https://reader034.fdocuments.us/reader034/viewer/2022042606/5f7f4a35fc5c7a39192a1a8c/html5/thumbnails/3.jpg)
How can GPU computing impact neurologic disease?
A longer story than you might think
![Page 4: How GPU Computing can Accelerate the Treatment of ... · PASCAL-VOC Test set Object detection on PASCAL-VOC Test set Classification on ImageNet ‘val’ set Effect of pre-training](https://reader034.fdocuments.us/reader034/viewer/2022042606/5f7f4a35fc5c7a39192a1a8c/html5/thumbnails/4.jpg)
3 Stories Enabling Neurosurgery Applications● Computing Power → Radiation Planning● Computing Localization → Intraoperative Applications● Computing Density → Medical ML/DL
Basically, “what happened to enable us to build department computing resources for AI that really work?”
And then, what does that look like?
![Page 5: How GPU Computing can Accelerate the Treatment of ... · PASCAL-VOC Test set Object detection on PASCAL-VOC Test set Classification on ImageNet ‘val’ set Effect of pre-training](https://reader034.fdocuments.us/reader034/viewer/2022042606/5f7f4a35fc5c7a39192a1a8c/html5/thumbnails/5.jpg)
Censor, Y., Altschuler, M. D. & Powlis, W. D. Appl. Math. Comput. 25, 57–87 (1988).
![Page 6: How GPU Computing can Accelerate the Treatment of ... · PASCAL-VOC Test set Object detection on PASCAL-VOC Test set Classification on ImageNet ‘val’ set Effect of pre-training](https://reader034.fdocuments.us/reader034/viewer/2022042606/5f7f4a35fc5c7a39192a1a8c/html5/thumbnails/6.jpg)
Censor, Y., Altschuler, M. D. & Powlis, W. D. Appl. Math. Comput. 25, 57–87 (1988).
![Page 7: How GPU Computing can Accelerate the Treatment of ... · PASCAL-VOC Test set Object detection on PASCAL-VOC Test set Classification on ImageNet ‘val’ set Effect of pre-training](https://reader034.fdocuments.us/reader034/viewer/2022042606/5f7f4a35fc5c7a39192a1a8c/html5/thumbnails/7.jpg)
Censor, Y., Altschuler, M. D. & Powlis, W. D. Appl. Math. Comput. 25, 57–87 (1988).https://www.brainlab.com/press-releases/brainlab-optimizes-planning-processes-algorithms-cranial-indications/
![Page 8: How GPU Computing can Accelerate the Treatment of ... · PASCAL-VOC Test set Object detection on PASCAL-VOC Test set Classification on ImageNet ‘val’ set Effect of pre-training](https://reader034.fdocuments.us/reader034/viewer/2022042606/5f7f4a35fc5c7a39192a1a8c/html5/thumbnails/8.jpg)
![Page 9: How GPU Computing can Accelerate the Treatment of ... · PASCAL-VOC Test set Object detection on PASCAL-VOC Test set Classification on ImageNet ‘val’ set Effect of pre-training](https://reader034.fdocuments.us/reader034/viewer/2022042606/5f7f4a35fc5c7a39192a1a8c/html5/thumbnails/9.jpg)
Fellner, F. A. J. Biomed. Sci. Eng. 9, 170 (2016)
![Page 10: How GPU Computing can Accelerate the Treatment of ... · PASCAL-VOC Test set Object detection on PASCAL-VOC Test set Classification on ImageNet ‘val’ set Effect of pre-training](https://reader034.fdocuments.us/reader034/viewer/2022042606/5f7f4a35fc5c7a39192a1a8c/html5/thumbnails/10.jpg)
![Page 11: How GPU Computing can Accelerate the Treatment of ... · PASCAL-VOC Test set Object detection on PASCAL-VOC Test set Classification on ImageNet ‘val’ set Effect of pre-training](https://reader034.fdocuments.us/reader034/viewer/2022042606/5f7f4a35fc5c7a39192a1a8c/html5/thumbnails/11.jpg)
![Page 12: How GPU Computing can Accelerate the Treatment of ... · PASCAL-VOC Test set Object detection on PASCAL-VOC Test set Classification on ImageNet ‘val’ set Effect of pre-training](https://reader034.fdocuments.us/reader034/viewer/2022042606/5f7f4a35fc5c7a39192a1a8c/html5/thumbnails/12.jpg)
Needs of academic, medical DL● Understand varied medical data needs● Mixed compute/data access patterns● Performance per dollar (financial constraints)● Access to appropriate storage that can handle imaging down to free text● Unified infrastructure, authentication and appropriate HIPAA privacy controls● Support for current and future generation computing paradigms
○ E.g., Docker, Container frameworks
![Page 13: How GPU Computing can Accelerate the Treatment of ... · PASCAL-VOC Test set Object detection on PASCAL-VOC Test set Classification on ImageNet ‘val’ set Effect of pre-training](https://reader034.fdocuments.us/reader034/viewer/2022042606/5f7f4a35fc5c7a39192a1a8c/html5/thumbnails/13.jpg)
Medical Imaging Data IS big dataConsider 1 megapixel, 8 bit detector (# in batch, z, x, y, # channels):
● Single slice / 2D image (1, 1, 1024, 1024, 1) = 1 Mb● 3D image with 100 slices (1, 100, 1024, 1024, 1) = 100 Mb● 1024 images/batch (1024, 100, 1024, 1024, 1) = 100 Gb
![Page 14: How GPU Computing can Accelerate the Treatment of ... · PASCAL-VOC Test set Object detection on PASCAL-VOC Test set Classification on ImageNet ‘val’ set Effect of pre-training](https://reader034.fdocuments.us/reader034/viewer/2022042606/5f7f4a35fc5c7a39192a1a8c/html5/thumbnails/14.jpg)
● Memory ● Precision ● Bandwidth
● Performance/$/Watt per application○ 2D Imaging○ 3D Volumetric Imaging○ NLP, RNN, Time Series○ Reinforcement Learning
● Comes down to:○ What’s your data?○ What’s your method?○ What’s your benchmark for performance?○ How rich are you and how much do you value your time?
![Page 15: How GPU Computing can Accelerate the Treatment of ... · PASCAL-VOC Test set Object detection on PASCAL-VOC Test set Classification on ImageNet ‘val’ set Effect of pre-training](https://reader034.fdocuments.us/reader034/viewer/2022042606/5f7f4a35fc5c7a39192a1a8c/html5/thumbnails/15.jpg)
http://timdettmers.com/2018/11/05/which-gpu-for-deep-learning/
![Page 16: How GPU Computing can Accelerate the Treatment of ... · PASCAL-VOC Test set Object detection on PASCAL-VOC Test set Classification on ImageNet ‘val’ set Effect of pre-training](https://reader034.fdocuments.us/reader034/viewer/2022042606/5f7f4a35fc5c7a39192a1a8c/html5/thumbnails/16.jpg)
Academic medical centers tend to start with what they know and evolve
![Page 17: How GPU Computing can Accelerate the Treatment of ... · PASCAL-VOC Test set Object detection on PASCAL-VOC Test set Classification on ImageNet ‘val’ set Effect of pre-training](https://reader034.fdocuments.us/reader034/viewer/2022042606/5f7f4a35fc5c7a39192a1a8c/html5/thumbnails/17.jpg)
Management● V1: Classic HPC Cluster
○ YP/NIS Authentication○ Manual Time Sharing○ NFS v3 XFS 20TB
● V2: Major Expansion, Not-So-Classic HPC Cluster○ Transition to Docker/Container Frameworks○ Manual Time Sharing○ Manual Authentication○ NFS v3 XFS 20TB + Local Flash/Scratch HDDs○ Flat/Volumetric Box Allocation to Specific Projects
![Page 18: How GPU Computing can Accelerate the Treatment of ... · PASCAL-VOC Test set Object detection on PASCAL-VOC Test set Classification on ImageNet ‘val’ set Effect of pre-training](https://reader034.fdocuments.us/reader034/viewer/2022042606/5f7f4a35fc5c7a39192a1a8c/html5/thumbnails/18.jpg)
![Page 19: How GPU Computing can Accelerate the Treatment of ... · PASCAL-VOC Test set Object detection on PASCAL-VOC Test set Classification on ImageNet ‘val’ set Effect of pre-training](https://reader034.fdocuments.us/reader034/viewer/2022042606/5f7f4a35fc5c7a39192a1a8c/html5/thumbnails/19.jpg)
Total Compute● “Flat” GPUs, Consumer GTX/RTX
○ Great bang for your buck, limited appropriateness for 3D volumetric work due to small amount of on-die memory (8-12GB)
○ 2 x GTX 1080 (FP32 8TF)○ 6 x GTX 1080 Ti (FT32 10TF)○ 2 x GTX 2080 Ti (FP32 14TF, 110TF w/ Tensor Cores)
● “Volumetric” GPUs, Mid-Level and Enterprise○ 3 - 10x Cost, ~double the memory○ 2 x Quadro P6000 (FP32 12TF, 24GB OD, FP64)○ 4 x RTX Titan (FP32 16TF, 130TF w/ Tensor Cores, 24GB OD, RP INT4/8 + FP16/64)○ 8 x Tesla V100 (FP32 16TF, 125TF w/ Tensor Cores, 32GB OD, RP INT4/8 + FP16/64)
● Total Tensor flops: 5.6PF + General Purpose FP32 @ 0.86PF
![Page 20: How GPU Computing can Accelerate the Treatment of ... · PASCAL-VOC Test set Object detection on PASCAL-VOC Test set Classification on ImageNet ‘val’ set Effect of pre-training](https://reader034.fdocuments.us/reader034/viewer/2022042606/5f7f4a35fc5c7a39192a1a8c/html5/thumbnails/20.jpg)
![Page 21: How GPU Computing can Accelerate the Treatment of ... · PASCAL-VOC Test set Object detection on PASCAL-VOC Test set Classification on ImageNet ‘val’ set Effect of pre-training](https://reader034.fdocuments.us/reader034/viewer/2022042606/5f7f4a35fc5c7a39192a1a8c/html5/thumbnails/21.jpg)
Management● V3: Next-Generation Containerized Cluster
○ Towards DeepOps○ NFS v4 288TB BTRFS RAID6 + HSs○ LDAP Unified Authentication (2 Factor + Sinai VPN)○ Role-Based Data Access Validation○ ContainerOS○ Kubernetes Docker Orchestration Framework○ Flat/Volumetric PXE Thin Nodes○ Managed Docker Containers for All Projects
![Page 22: How GPU Computing can Accelerate the Treatment of ... · PASCAL-VOC Test set Object detection on PASCAL-VOC Test set Classification on ImageNet ‘val’ set Effect of pre-training](https://reader034.fdocuments.us/reader034/viewer/2022042606/5f7f4a35fc5c7a39192a1a8c/html5/thumbnails/22.jpg)
![Page 23: How GPU Computing can Accelerate the Treatment of ... · PASCAL-VOC Test set Object detection on PASCAL-VOC Test set Classification on ImageNet ‘val’ set Effect of pre-training](https://reader034.fdocuments.us/reader034/viewer/2022042606/5f7f4a35fc5c7a39192a1a8c/html5/thumbnails/23.jpg)
How can machine learning (on GPUs) impact neurological disease?
A universe of new applications
![Page 24: How GPU Computing can Accelerate the Treatment of ... · PASCAL-VOC Test set Object detection on PASCAL-VOC Test set Classification on ImageNet ‘val’ set Effect of pre-training](https://reader034.fdocuments.us/reader034/viewer/2022042606/5f7f4a35fc5c7a39192a1a8c/html5/thumbnails/24.jpg)
Assessments in the Neuro-ICU
Davoudi, A. et al. The Intelligent ICU Pilot Study: Using Artificial Intelligence Technology for Autonomous Patient Monitoring. arXiv [cs.HC] (2018).
![Page 25: How GPU Computing can Accelerate the Treatment of ... · PASCAL-VOC Test set Object detection on PASCAL-VOC Test set Classification on ImageNet ‘val’ set Effect of pre-training](https://reader034.fdocuments.us/reader034/viewer/2022042606/5f7f4a35fc5c7a39192a1a8c/html5/thumbnails/25.jpg)
Davoudi, A. et al. The Intelligent ICU Pilot Study: Using Artificial Intelligence Technology for Autonomous Patient Monitoring. arXiv [cs.HC] (2018).
![Page 26: How GPU Computing can Accelerate the Treatment of ... · PASCAL-VOC Test set Object detection on PASCAL-VOC Test set Classification on ImageNet ‘val’ set Effect of pre-training](https://reader034.fdocuments.us/reader034/viewer/2022042606/5f7f4a35fc5c7a39192a1a8c/html5/thumbnails/26.jpg)
Davoudi, A. et al. The Intelligent ICU Pilot Study: Using Artificial Intelligence Technology for Autonomous Patient Monitoring. arXiv [cs.HC] (2018).
![Page 27: How GPU Computing can Accelerate the Treatment of ... · PASCAL-VOC Test set Object detection on PASCAL-VOC Test set Classification on ImageNet ‘val’ set Effect of pre-training](https://reader034.fdocuments.us/reader034/viewer/2022042606/5f7f4a35fc5c7a39192a1a8c/html5/thumbnails/27.jpg)
Convolutional Neural Network Approaches to Brain Imaging
![Page 28: How GPU Computing can Accelerate the Treatment of ... · PASCAL-VOC Test set Object detection on PASCAL-VOC Test set Classification on ImageNet ‘val’ set Effect of pre-training](https://reader034.fdocuments.us/reader034/viewer/2022042606/5f7f4a35fc5c7a39192a1a8c/html5/thumbnails/28.jpg)
Classification and Localization● Input: N classes + BBox (x,y,w,h)● Output: Class K where K is in N + (xp,yp,wp,hp)● Performance Metrics: Accuracy + Jaccard similarity (or Dice)
conv layers+/- pooling
+/- fully conn layers
CORGI
Final conv layer
Softmax LOSS: CCE
(xp,yp,wp,hp)
LOSS: L2
![Page 29: How GPU Computing can Accelerate the Treatment of ... · PASCAL-VOC Test set Object detection on PASCAL-VOC Test set Classification on ImageNet ‘val’ set Effect of pre-training](https://reader034.fdocuments.us/reader034/viewer/2022042606/5f7f4a35fc5c7a39192a1a8c/html5/thumbnails/29.jpg)
Segmentation and Classification
conv layers+/- pooling
+/- fully conn layers
CORGI
Final conv layer
Softmax LOSS: CCE
![Page 30: How GPU Computing can Accelerate the Treatment of ... · PASCAL-VOC Test set Object detection on PASCAL-VOC Test set Classification on ImageNet ‘val’ set Effect of pre-training](https://reader034.fdocuments.us/reader034/viewer/2022042606/5f7f4a35fc5c7a39192a1a8c/html5/thumbnails/30.jpg)
Brain Biopsies
Zhou, M. et al. Radiomics in Brain Tumor: Image Assessment, Quantitative Feature Descriptors and Machine-learning Approaches. AJNR Am. J. Neuroradiol. 39, 208 (2018).
![Page 31: How GPU Computing can Accelerate the Treatment of ... · PASCAL-VOC Test set Object detection on PASCAL-VOC Test set Classification on ImageNet ‘val’ set Effect of pre-training](https://reader034.fdocuments.us/reader034/viewer/2022042606/5f7f4a35fc5c7a39192a1a8c/html5/thumbnails/31.jpg)
Brain Biopsies
Chang, P. et al. Deep-Learning Convolutional Neural Networks Accurately Classify Genetic Mutations in Gliomas. AJNR Am. J. Neuroradiol. (2018). doi:10.3174/ajnr.A5667
![Page 32: How GPU Computing can Accelerate the Treatment of ... · PASCAL-VOC Test set Object detection on PASCAL-VOC Test set Classification on ImageNet ‘val’ set Effect of pre-training](https://reader034.fdocuments.us/reader034/viewer/2022042606/5f7f4a35fc5c7a39192a1a8c/html5/thumbnails/32.jpg)
Weak Supervision
![Page 33: How GPU Computing can Accelerate the Treatment of ... · PASCAL-VOC Test set Object detection on PASCAL-VOC Test set Classification on ImageNet ‘val’ set Effect of pre-training](https://reader034.fdocuments.us/reader034/viewer/2022042606/5f7f4a35fc5c7a39192a1a8c/html5/thumbnails/33.jpg)
![Page 34: How GPU Computing can Accelerate the Treatment of ... · PASCAL-VOC Test set Object detection on PASCAL-VOC Test set Classification on ImageNet ‘val’ set Effect of pre-training](https://reader034.fdocuments.us/reader034/viewer/2022042606/5f7f4a35fc5c7a39192a1a8c/html5/thumbnails/34.jpg)
Two Kinds of Labels
Gold Standard LabelsGround Truth
Silver Standard LabelsNoisy Labels
![Page 35: How GPU Computing can Accelerate the Treatment of ... · PASCAL-VOC Test set Object detection on PASCAL-VOC Test set Classification on ImageNet ‘val’ set Effect of pre-training](https://reader034.fdocuments.us/reader034/viewer/2022042606/5f7f4a35fc5c7a39192a1a8c/html5/thumbnails/35.jpg)
Are Medical GT Labels Fool’s Gold? ● Medical labels can be challenging
with low IRR○ Google Retinopathy dataset =
55.4% ○ IRR and 70.1% agreement
between each expert and her/himself at a later time point!
● Can average labels using EM.● However, average of modeled raters
may outperform model of average raters.
● Guan et al. 2017 had 1.97% decrease in test loss
Guan et al. 2017 - Who Said What - Modeling Individual Labelers Improves ClassificationWhitehill et al. 2009 - Whose Vote Should Count More - Optimal Integration of Labels from Labelers of Unknown Expertise
![Page 36: How GPU Computing can Accelerate the Treatment of ... · PASCAL-VOC Test set Object detection on PASCAL-VOC Test set Classification on ImageNet ‘val’ set Effect of pre-training](https://reader034.fdocuments.us/reader034/viewer/2022042606/5f7f4a35fc5c7a39192a1a8c/html5/thumbnails/36.jpg)
Weak Supervision with Generated Silver LabelsSolution? Accept noise in our label set.
Alex Ratner, Stephen Bach and Chris Ré - Snorkel Blog
![Page 37: How GPU Computing can Accelerate the Treatment of ... · PASCAL-VOC Test set Object detection on PASCAL-VOC Test set Classification on ImageNet ‘val’ set Effect of pre-training](https://reader034.fdocuments.us/reader034/viewer/2022042606/5f7f4a35fc5c7a39192a1a8c/html5/thumbnails/37.jpg)
The Unreasonable Effectiveness of Big Data with Silver Labels
C Sun, et al. Revisiting Unreasonable Effectiveness of Data in Deep Learning Era - arXiv 2017
But does this work? Consider the following trends in computer vision with ImageNet….
What if we had a dataset 300x ImageNet’s size with noisy labels?
![Page 38: How GPU Computing can Accelerate the Treatment of ... · PASCAL-VOC Test set Object detection on PASCAL-VOC Test set Classification on ImageNet ‘val’ set Effect of pre-training](https://reader034.fdocuments.us/reader034/viewer/2022042606/5f7f4a35fc5c7a39192a1a8c/html5/thumbnails/38.jpg)
The Unreasonable Effectiveness of Big DataSemantic segmentation on
PASCAL-VOC Test set
Object detection on PASCAL-VOC Test set
Classification on ImageNet ‘val’ set
Effect of pre-training ResNet-101 on JFT-300M’s silver labels
C Sun, et al. Revisiting Unreasonable Effectiveness of Data in Deep Learning Era - arXiv 2017
![Page 39: How GPU Computing can Accelerate the Treatment of ... · PASCAL-VOC Test set Object detection on PASCAL-VOC Test set Classification on ImageNet ‘val’ set Effect of pre-training](https://reader034.fdocuments.us/reader034/viewer/2022042606/5f7f4a35fc5c7a39192a1a8c/html5/thumbnails/39.jpg)
Application to Acute Neurologic Events
Titano, J. J. et al. Automated deep-neural-network surveillance of cranial images for acute neurologic events. Nat. Med. (2018). doi:10.1038/s41591-018-0147-y
![Page 40: How GPU Computing can Accelerate the Treatment of ... · PASCAL-VOC Test set Object detection on PASCAL-VOC Test set Classification on ImageNet ‘val’ set Effect of pre-training](https://reader034.fdocuments.us/reader034/viewer/2022042606/5f7f4a35fc5c7a39192a1a8c/html5/thumbnails/40.jpg)
Faster Interpretation of Imaging
Titano, J. J. et al. Automated deep-neural-network surveillance of cranial images for acute neurologic events. Nat. Med. (2018). doi:10.1038/s41591-018-0147-y
![Page 41: How GPU Computing can Accelerate the Treatment of ... · PASCAL-VOC Test set Object detection on PASCAL-VOC Test set Classification on ImageNet ‘val’ set Effect of pre-training](https://reader034.fdocuments.us/reader034/viewer/2022042606/5f7f4a35fc5c7a39192a1a8c/html5/thumbnails/41.jpg)
Faster Interpretation of Imaging
Titano, J. J. et al. Automated deep-neural-network surveillance of cranial images for acute neurologic events. Nat. Med. (2018). doi:10.1038/s41591-018-0147-y
![Page 42: How GPU Computing can Accelerate the Treatment of ... · PASCAL-VOC Test set Object detection on PASCAL-VOC Test set Classification on ImageNet ‘val’ set Effect of pre-training](https://reader034.fdocuments.us/reader034/viewer/2022042606/5f7f4a35fc5c7a39192a1a8c/html5/thumbnails/42.jpg)
Disclaimer #1: Generalization of deep models is not guaranteed
Zhang, C., Bengio, S., Hardt, M., Recht, B. & Vinyals, O. Understanding deep learning requires rethinking generalization. arXiv [cs.LG] (2016).
![Page 43: How GPU Computing can Accelerate the Treatment of ... · PASCAL-VOC Test set Object detection on PASCAL-VOC Test set Classification on ImageNet ‘val’ set Effect of pre-training](https://reader034.fdocuments.us/reader034/viewer/2022042606/5f7f4a35fc5c7a39192a1a8c/html5/thumbnails/43.jpg)
Disclaimer #2: Weak Classifiers are Easily Distracted
('bucket', 0.43788964), ('tub', 0.13390972), ('caldron', 0.11801116)
Average Precision (AP) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.900Average Precision (AP) @[ IoU=0.50 | area= all | maxDets=100 ] = 1.000Average Precision (AP) @[ IoU=0.75 | area= all | maxDets=100 ] = 1.000
![Page 44: How GPU Computing can Accelerate the Treatment of ... · PASCAL-VOC Test set Object detection on PASCAL-VOC Test set Classification on ImageNet ‘val’ set Effect of pre-training](https://reader034.fdocuments.us/reader034/viewer/2022042606/5f7f4a35fc5c7a39192a1a8c/html5/thumbnails/44.jpg)
Disclaimer #2: Weak Classifiers are Easily Distracted
![Page 45: How GPU Computing can Accelerate the Treatment of ... · PASCAL-VOC Test set Object detection on PASCAL-VOC Test set Classification on ImageNet ‘val’ set Effect of pre-training](https://reader034.fdocuments.us/reader034/viewer/2022042606/5f7f4a35fc5c7a39192a1a8c/html5/thumbnails/45.jpg)
Disclaimer #3: Data is Everything
![Page 46: How GPU Computing can Accelerate the Treatment of ... · PASCAL-VOC Test set Object detection on PASCAL-VOC Test set Classification on ImageNet ‘val’ set Effect of pre-training](https://reader034.fdocuments.us/reader034/viewer/2022042606/5f7f4a35fc5c7a39192a1a8c/html5/thumbnails/46.jpg)
Disclaimer #4: Medical Data Paid for in Human LivesWe are going to need more training data...
![Page 47: How GPU Computing can Accelerate the Treatment of ... · PASCAL-VOC Test set Object detection on PASCAL-VOC Test set Classification on ImageNet ‘val’ set Effect of pre-training](https://reader034.fdocuments.us/reader034/viewer/2022042606/5f7f4a35fc5c7a39192a1a8c/html5/thumbnails/47.jpg)
MSHS MEDICAL A.I. CONSORTIUMOrthopedicsRadiology
Samuel Cho, MDAssociate Professor, Department of Orthopedics and Neurosurgery
Jun Kim, MDHouseofficer, Department of Orthopedic Surgery
Komal Srivastava, BAGrant Specialist, Department of Orthopedic Surgery
Eric Karl Oermann, MDInstructor, Department of Neurosurgery
Anthony Costa, PhDAssistant Professor, Department of NeurosurgeryDirector, Sinai BioDesign
Joshua B Bederson, MDChairman, Department of Neurosurgery
Holly Oemke, BAProgram Manager, Sinai Biodesign
Margaret Pain, MDHouseofficer, Department of Neurosurgery
Raj Shrivastava, MDAssociate Professor, Department of Neurosurgery
John Caridi, MDAssistant Professor, Department of Neurosurgery
Neha Dangayach, MDAssistant Professor, Department of NeurosurgeryResearch co-director for ICCM
Neurological SurgeryAISINAI
COLLABORATORS:Merck:Joseph Lehar, PhDDirector of Computational Biology
Hammerlab: Alex Rubinsteyn, PhDPostdoc, GGS
Intel:Peter Tang, PhDSenior Fellow
Google: Marcus Badgeley, MEngPhD student, Google / VerilyMedical student, ISMMSISMMS
Fred Kwon, MSEMD/PhD student
Martin Kang, BSMedical Student
Deepak Kaji, BSMD/PhD student
Varun Arvind, BSMD/PhD student
Alice Fan, MDAssistant Professor of Oncology
Viola Chen, MDFellow, Department of Oncology
Joseph Titano, MDFellow, Department of Radiology
Javin Schefflein, MDHouseofficer, Department of Radiology
Burton Drayer, MDChairman, Department of Radiology
Brett Marinelli, MDHouseofficer, Department of Radiology
Nathaniel Swinburne, MDHouseofficer, Department of Radiology
Andres Su, MDHouseofficer, Department of Radiology
Michael Cai, MDHouseofficer, Department of Radiology
Sonam Sharma, MDAssistant Professor, Radiation Oncology
Radiation Oncology
Zahi Fayad, PhDDirector MSHS TMIIIDavid Mendelson, MDDirector of Informatics