{"title":"深度学习和人工智能可以用估计置信度解决测量问题","authors":"Uğurcan Akyüz","doi":"10.51843/wsproceedings.2020.18","DOIUrl":null,"url":null,"abstract":"In the past, like many companies, we tried to automate the data collection of handheld meters and other devices using Optical Character Recognition (OCR) technology, only to learn OCR technology has its limitations; we discovered any change in position, lighting, angle and even glare would throw off the OCR, resulting in bad numbers. So we changed direction and switched to Artificial Intelligence (AI) with Deep Learning algorithms. Our goal was to implement a “learn as you go,” AI-assisted solution that will learn to read a handheld meter as good, or better than, a human. The continual learning/training would train the AI to read measurements at any angle, in most lighting conditions. The trained AI would even be smart enough to understand scaled values based on the prefixes. Over time, with a much larger data set, the AI would be able to read just about any new display.","PeriodicalId":422993,"journal":{"name":"NCSL International Workshop & Symposium Conference Proceedings 2020","volume":"57 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"1900-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Deep Learning & Artificial Intelligence can Solve Measurement Problems with Estimated Confidence \",\"authors\":\"Uğurcan Akyüz\",\"doi\":\"10.51843/wsproceedings.2020.18\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"In the past, like many companies, we tried to automate the data collection of handheld meters and other devices using Optical Character Recognition (OCR) technology, only to learn OCR technology has its limitations; we discovered any change in position, lighting, angle and even glare would throw off the OCR, resulting in bad numbers. So we changed direction and switched to Artificial Intelligence (AI) with Deep Learning algorithms. Our goal was to implement a “learn as you go,” AI-assisted solution that will learn to read a handheld meter as good, or better than, a human. The continual learning/training would train the AI to read measurements at any angle, in most lighting conditions. The trained AI would even be smart enough to understand scaled values based on the prefixes. Over time, with a much larger data set, the AI would be able to read just about any new display.\",\"PeriodicalId\":422993,\"journal\":{\"name\":\"NCSL International Workshop & Symposium Conference Proceedings 2020\",\"volume\":\"57 1\",\"pages\":\"0\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"1900-01-01\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"NCSL International Workshop & Symposium Conference Proceedings 2020\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.51843/wsproceedings.2020.18\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"NCSL International Workshop & Symposium Conference Proceedings 2020","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.51843/wsproceedings.2020.18","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
Deep Learning & Artificial Intelligence can Solve Measurement Problems with Estimated Confidence
In the past, like many companies, we tried to automate the data collection of handheld meters and other devices using Optical Character Recognition (OCR) technology, only to learn OCR technology has its limitations; we discovered any change in position, lighting, angle and even glare would throw off the OCR, resulting in bad numbers. So we changed direction and switched to Artificial Intelligence (AI) with Deep Learning algorithms. Our goal was to implement a “learn as you go,” AI-assisted solution that will learn to read a handheld meter as good, or better than, a human. The continual learning/training would train the AI to read measurements at any angle, in most lighting conditions. The trained AI would even be smart enough to understand scaled values based on the prefixes. Over time, with a much larger data set, the AI would be able to read just about any new display.