Proceedings of the First International Conference on AI-ML Systems最新文献

Compression of Deep Learning Models for NLP NLP中深度学习模型的压缩

Proceedings of the First International Conference on AI-ML Systems Pub Date : 2021-10-21 DOI: 10.1145/3486001.3486247

Manish Gupta

引用次数: 0

Optimizing Convolutions for an Inference Accelerator: Case Study: Intel’s NNP-I 1000 DL Compute Grid 为推理加速器优化卷积:案例研究:英特尔的NNP-I 1000 DL计算网格

Proceedings of the First International Conference on AI-ML Systems Pub Date : 2021-10-21 DOI: 10.1145/3486001.3486239

Durgadoss, Kausik Maiti, Sanju C Sudhakaran, Isha Agarwal, Kartik Podugu, Pavan Kumar, Jitender Patil, A. Chawla

引用次数: 0

CryptInfer: Enabling Encrypted Inference on Skin Lesion Images for Melanoma Detection CryptInfer:对皮肤病变图像进行加密推断，用于黑色素瘤检测

Proceedings of the First International Conference on AI-ML Systems Pub Date : 2021-10-21 DOI: 10.1145/3486001.3486233

Nayna Jain, Karthik Nandakumar, N. Ratha, Sharath Pankanti, U. Kumar

{"title":"CryptInfer: Enabling Encrypted Inference on Skin Lesion Images for Melanoma Detection","authors":"Nayna Jain, Karthik Nandakumar, N. Ratha, Sharath Pankanti, U. Kumar","doi":"10.1145/3486001.3486233","DOIUrl":"https://doi.org/10.1145/3486001.3486233","url":null,"abstract":"Deep learning models such as Convolutional Neural Networks (CNNs) have shown the potential to classify medical images for accurate diagnosis. These techniques will face regulatory compliance challenges related to privacy of user data, especially when they are deployed as a service on a cloud platform. Fully Homomorphic Encryption (FHE) can enable CNN inference on encrypted data and help mitigate such concerns. However, encrypted CNN inference faces the fundamental challenge of optimizing the computations to achieve an acceptable trade-off between accuracy and practical computational feasibility. Current approaches for encrypted CNN inference demonstrate feasibility typically on smaller images (e.g., MNIST and CIFAR-10 datasets) and shallow neural networks. This work is the first to show encrypted inference results on a real-world dataset for melanoma detection with large-sized images of skin lesions based on the Cheon-Kim-Kim-Song (CKKS) encryption scheme available in the open-source HElib library. The practical challenges related to encrypted inference are first analyzed and inference experiments are conducted on encrypted MNIST images to evaluate different optimization strategies and their role in determining the throughput and latency of the inference process. Using these insights, a modified LeNet-like architecture is designed and implemented to achieve the end goal of enabling encrypted inference on melanoma dataset. The results demonstrate that 80% classification accuracy can be achieved on encrypted skin lesion images (security of 106 bits) with a latency of 51 seconds for single image inference and a throughput of 18,000 images per hour for batched inference, which shows that privacy-preserving machine learning as a service (MLaaS) based on encrypted data is indeed practically feasible.","PeriodicalId":266754,"journal":{"name":"Proceedings of the First International Conference on AI-ML Systems","volume":"2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-10-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127937227","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 1

Real-time Social Distancing Detection System with Auto-calibration using Pose Information 使用姿态信息自动校准的实时社会距离检测系统

Proceedings of the First International Conference on AI-ML Systems Pub Date : 2021-10-21 DOI: 10.1145/3486001.3486245

Gaku Nakano, Shoji Nishimura

引用次数: 7

Bridging between Data Science and Performance Analysis: Tracing of Jupyter Notebooks 数据科学和性能分析之间的桥梁:Jupyter笔记本的跟踪

Proceedings of the First International Conference on AI-ML Systems Pub Date : 2021-10-21 DOI: 10.1145/3486001.3486249

Elias Werner, Lalith Manjunath, Jan Frenzel, Sunna Torge

引用次数: 1

Word-level beam search decoding and correction algorithm (WLBS) for end-to-end ASR 端到端ASR的字级波束搜索译码校正算法

Proceedings of the First International Conference on AI-ML Systems Pub Date : 2021-10-21 DOI: 10.1145/3486001.3486223

S. Zitha, Prabhakar Venkata Tamma

引用次数: 0

A Deep Learning framework for Interoperable Machine Learning 面向互操作机器学习的深度学习框架

Proceedings of the First International Conference on AI-ML Systems Pub Date : 2021-10-21 DOI: 10.1145/3486001.3486243

Shakkeel Ahmed, Prakash Bisht, Ravi Mula, S. Dhavala

引用次数: 2

Stacked Sparse Autoencoder and Machine Learning Based Anxiety Classification Using EEG Signals 基于脑电信号的堆叠稀疏自编码器和机器学习焦虑分类

Proceedings of the First International Conference on AI-ML Systems Pub Date : 2021-10-21 DOI: 10.1145/3486001.3486227

S. Shikha, Manan Agrawal, Mohd Anwar, Divyashikha Sethia

引用次数: 5

InuAid Chest X-rays AI Product: Respiratory diseases detection using Artificial Intelligence to significantly improve productivity and quality of diagnosis InuAid胸部x光人工智能产品:利用人工智能进行呼吸系统疾病检测，显著提高诊断效率和质量

Proceedings of the First International Conference on AI-ML Systems Pub Date : 2021-10-21 DOI: 10.1145/3486001.3486242

Prem Kumar S. R., Vikas Kumar Anand, S. Senapati, A. Murali

{"title":"InuAid Chest X-rays AI Product: Respiratory diseases detection using Artificial Intelligence to significantly improve productivity and quality of diagnosis","authors":"Prem Kumar S. R., Vikas Kumar Anand, S. Senapati, A. Murali","doi":"10.1145/3486001.3486242","DOIUrl":"https://doi.org/10.1145/3486001.3486242","url":null,"abstract":"The spread of Covid-19 virus around the world has taken many lives, quarantined people and shattered many industries. Due to high transmissibility of the virus and its silent incubation period in human beings, detection of the virus plays an important role to control its spread and to plan diagnostic and preventive measures. Laboratory tests such as Polymerase Chain Reaction (PCR) take more time and hence there is a need for rapid and accurate diagnostic methods to detect the virus to prevent its spread and combat it. Today PCR tests were used for diagnosing purposes and the chest x-ray was only used as the follow up of patients, hence these studies on the chest x-rays of patients with Covid-19 pneumonia or any other disease are still limited to the literature and must be improved in the future. In this project, the goal is to build an application for healthcare workers to monitor the health of lungs using the chest x-ray images of patients. The algorithm must be very accurate because it deals with the lives of people. Here we used computer vision and deep learning techniques in this project. The focus is to classify chest x-ray images and segment the abnormal region and to get more insights on the images from the available datasets. The diagnostic accuracy is the challenging part and to increase the detection efficiency due to the limited open-source data available. The data was collected from the internet. On classification, the trained model was able to achieve 93.10% accuracy and F1 Score of 0.93 after using transfer learning technique with pneumonia images. On segmentation, the Intersection Over Union value was found to be 0.91 on the validation data.","PeriodicalId":266754,"journal":{"name":"Proceedings of the First International Conference on AI-ML Systems","volume":"3 6","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-10-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114132816","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Leveraging Prediction Confidence For Versatile Optimizations to CNNs 利用预测置信度对cnn进行通用优化

Proceedings of the First International Conference on AI-ML Systems Pub Date : 2021-10-21 DOI: 10.1145/3486001.3486222

Saksham Sharma, Vanshika V Bhargava, Aditya Singh, K. Bhardwaj, Sparsh Mittal

{"title":"Leveraging Prediction Confidence For Versatile Optimizations to CNNs","authors":"Saksham Sharma, Vanshika V Bhargava, Aditya Singh, K. Bhardwaj, Sparsh Mittal","doi":"10.1145/3486001.3486222","DOIUrl":"https://doi.org/10.1145/3486001.3486222","url":null,"abstract":"Modern convolutional neural networks (CNNs) incur huge computational and energy overheads. In this paper, we propose two techniques for inferring the confidence in the correctness of a prediction in the early layers of a CNN. The first technique uses a statistical approach, whereas the second technique requires retraining. We argue that prediction confidence estimation can enable diverse optimizations to CNNs. We demonstrate two optimizations. First, we predict selected images in early layers. This is possible because in a dataset, many images are easy to predict and they can be predicted in the early layers of a CNN. This reduces the average computation count at the cost of accuracy and parameter count. Second, we propose predicting only selected images for which the prediction-confidence is high. This reduces the coverage; however, the accuracy on the images that are predicted is higher. Our results with VGG16 and ResNet50 CNNs on the Caltech256 dataset show that our techniques are effective. For example, for ResNet, our first technique reduces the accuracy from 71.6% to 69.8% while reducing the computations by 14%. Similarly, with the second technique, on reducing the coverage from 100% to 90%, the accuracy is increased from 71.6% to 75.6%. Keywords: computer vision, CNN, approximate computing, accuracy-coverage tradeoff, prediction confidence","PeriodicalId":266754,"journal":{"name":"Proceedings of the First International Conference on AI-ML Systems","volume":"118 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-10-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114371233","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0