php-ai/php-ml
In just a matter of weeks, the world around all of us has changed. Personally, when I wake up every morning now, my first thought goes to my elderly grandparents in their mid -70s living with chronic health conditions. As the coronavirus rampages around the world, changing societies in unprecedented ways and claiming lives as we speak, I am sure all of you are in one way or the other concerned about your loved ones and your community. Many of us are sharing a sense of vulnerability physically, socially or mentally.
在短短几周内，我们所有人的世界都发生了变化。 我个人每天早晨醒来时，首先想到的是70岁左右，患有慢性疾病的年迈的祖父母。 随着冠状病毒在世界范围内肆虐，以前所未有的方式改变社会并夺取我们的生命，我相信你们所有人都在某种程度上关注您的亲人和社区。 我们中的许多人在身体，社交或精神上都分享一种脆弱感。
In these hard times, I am especially inspired by the eagerness and diligence of scientists, mathematicians, engineers, social scientists and especially healthcare workers around the world who are coming together to combat this pandemic. A crisis like the COVID-19 outbreak presents challenges that touch many areas of life and span many fields of expertise. They don’t lend themselves to easy solutions or quick fixes. But, however challenging or daunting, the journey towards solutions must begin.
在这些困难时期，我尤其受到世界各地科学家，数学家，工程师，社会科学家，尤其是世界各地为抗击这一流行病而齐心协力的热情和勤奋的鼓舞。 像COVID-19疫情一样的危机带来了挑战，这些挑战涉及生活的许多领域，涉及许多专业领域。 他们不适合简单的解决方案或快速修复。 但是，无论挑战或艰巨，都必须开始寻求解决方案。
In December 2019, Bluedot, an Al platform which scans hundreds and thousands of foreign-language news reports, animal and plant disease networks and other reputable sources has detected a cluster of pneumonia-like illness in Wuhan, China. This system alerted the Blue Dots experts, who identified it as parallel to the deadly SARS outbreak in 2003. Even before scientists realized the gravity of the situation, and Al system has predicted the upcoming peril. Using global airline ticketing data, this platform has also predicted that the virus would jump from Wuhan to Seoul, Tokyo, Bangkok and Taipei in the upcoming days which turned out to be true. This shows that the importance of Al use during this crisis; advancements in data analytics, natural language processing (NLP), Machine Learning, Deep Learning, Computer Vision and Chatbots have not only helped to identify, track, forecast the pandemic, but also to support the medical staff by helping them in diagnosing the patients. Now, let’s look at a high level on how AI is helping the healthcare workers and doctors around the world.
2019年12月，一个可以扫描成千上万种外语新闻报道，动植物疾病网络和其他知名消息源的Al平台Bluedot在中国武汉发现了一系列肺炎样疾病。 该系统向蓝点专家发出警报，他们将其识别为与2003年致命的SARS爆发平行。甚至在科学家们意识到情况的严重性之前，Al系统就预言了即将来临的危险。 利用全球机票数据，该平台还预测该病毒将在未来几天内从武汉跳到汉城，东京，曼谷和台北，事实证明这是事实。 这表明在这次危机中使用铝的重要性。 数据分析，自然语言处理(NLP)，机器学习，深度学习，计算机视觉和聊天机器人的进步不仅有助于识别，跟踪和预测大流行病，而且还通过帮助医务人员诊断患者来支持他们。 现在，让我们从较高的角度看一下AI如何帮助世界各地的医护人员和医生。
用于疫苗研究的AI： (AI For Vaccine Research:)
Problem — Drug discovery is noted for complex, time-taking and expensive process which involves many trial and error methods requiring the assiduous endeavour of the world’s brightest minds. As the world is trying hard to fight against coronavirus, Al is offering new hope that a vaccine can be developed quicker than the time taken by traditional methods.
问题—药物发现被认为是复杂，耗时且昂贵的过程，其中涉及许多尝试和错误方法，需要世界上最聪明的人孜孜以求。 随着世界正努力与冠状病毒作斗争，A1带来了新希望，即疫苗的开发速度可以比传统方法更快。
Solution — Researchers are trying to discover possible drug combinations which can destroy the SARS COV-2 virus which is the name of the virus causing COVID 19 catastrophe. There are five Al centred companies which are accelerating this drug development. Among them, Deargen has made a valuable prediction. They published a paper using the deep learning-based model called MT-DTI which uses simplified chemical sequences to predict how strong a molecule of interest will bind to a target protein. This model predicted that of FDA approved antiviral drugs, the HIV medication atazanavir and Remdesivir can efficiently block a prominent protein on the outside of SARS-Cov-2.
解决方案—研究人员正在尝试发现可能破坏SARS COV-2病毒的可能药物组合，SARS COV-2病毒是引起COVID 19灾难的病毒的名称。 有五家以铝为中心的公司正在加速这种药物的开发。 其中， 迪尔根(Deargen )做出了有价值的预测。 他们使用基于深度学习的模型MT-DTI发表了一篇论文，该模型使用简化的化学序列来预测目标分子与目标蛋白结合的强度。 该模型预测，在FDA批准的抗病毒药物中，HIV药物atazanavir和Remdesivir可以有效地阻断SARS-Cov-2外部的突出蛋白。
Today, as predicted, Remdesivir is being touted as a possible treatment for COVID-19, but it’s still undergoing tests and not advertised as a cure. In a study conducted by Gilead (an American company who owns the Remdesivir patent globally) conducted a study on 229 patients in the United States on patients of different ethnicity to test the safety and efficacy of the drug. The results are as below in the table.
如今，正如预测的那样，瑞姆昔韦被吹捧为可能用于治疗COVID-19的方法，但它仍在接受测试，而没有广告宣传其治疗方法。 由吉利德(Gilead)(一家在全球拥有Remdesivir专利的美国公司)进行的一项研究针对美国229名不同种族的患者进行了研究，以测试该药物的安全性和有效性。 结果如下表所示。
Companies like Insilico Medicine, SRI Biosciences and Iktos, Benevolent Al are also scouting out Deep Learning models to predict old and new drugs that might cure COVID-19. Although a lot of research is happening in the industry to find a panacea, a systematic approach must be followed and a proper collaboration of bright minds in Al will better lead to a positive outcome.
像Insilico Medicine，SRI Biosciences和Iktos，Benevolent Al等公司也正在研究深度学习模型，以预测可能治愈COVID-19的新旧药物。 尽管业界正在进行大量研究来寻找灵丹妙药，但必须遵循系统的方法，Al中聪明才智的适当合作将更好地带来积极的结果。
用于快速诊断的AI： (AI for Quick Diagnosis:)
Making a quick diagnosis based on the symptoms of a patient can be a pretty tricky and hard process to deal with. Albeit, finding out the positive cases quickly and taking important remedies is an essential effort to fight against COVID 19. Al can help the doctors to quickly diagnose a patient. One method China is using is Al-based multisensory technology which is deployed in public places such as airports. This system can screen up to 200 people per minute and can detect their temperature within a range of 0.5 degrees Celsius. This allows the airport authorities to quickly isolate a person who is suspected of having a fever. However, this might just be a method to separate high fever persons, below are the prominent methods which are being used in the hospitals right now.
根据患者的症状进行快速诊断可能是一个非常棘手且困难的过程。 尽管Swift发现阳性病例并采取重要的补救措施是对抗COVID 19的必不可少的努力。Al可以帮助医生快速诊断患者。 中国使用的一种方法是在机场等公共场所部署的基于Al的多传感器技术。 该系统每分钟最多可以筛查200个人，并且可以检测他们在0.5摄氏度范围内的温度。 这使机场当局可以Swift隔离怀疑发烧的人。 但是，这可能只是区分高烧人群的一种方法，以下是当前医院中使用的主要方法。
1.冠状病毒诊断卷 (1. CORONA VIRUS DIAGNOSTIC CHATBOT)
Problem- Hospitals are being overburdened with many people and cases. People are unnecessarily panicking and approaching hospitals to get themselves tested. This causes serious problems like long waiting time to take the test at hospitals and unavailability of test kits for those in need. There is a risk that additional foot traffic can have increasing contamination between healthy and infected persons. Because of this, the healthcare systems are being overwhelmed and this is causing a delay for persons who are really affected with COVID-19 and need a quick diagnosis.
问题医院的人口和病例过多。 人们不必要地恐慌和接近医院进行自我测试。 这会导致严重的问题，例如等待医院进行测试的漫长等待时间以及无法为有需要的人提供测试套件。 存在额外的步行流量可能增加健康人和感染者之间的污染的风险。 因此，医疗保健系统不堪重负，这给真正受COVID-19影响并需要快速诊断的人造成了延误。
Solution- In order to tackle this, a company called Paginemediche is has leveraged the power of NLP(natural language processing) and developed a multilingual version of a Coronavirus chatbot which helps in triaging and diagnosing, symptoms checking and monitoring on the people who use it. This helps in providing reliable information and clear guidelines to people and can advise the individuals if they need hospital screening or self-isolation at their homes.
解决方案-为了解决这个问题，一家名为Paginemediche的公司利用了NLP(自然语言处理)的强大功能，并开发了多语言版本的冠状病毒聊天机器人，该机器人可帮助对其使用者进行分类和诊断，症状检查和监控。 。 这有助于为人们提供可靠的信息和明确的指导方针，并可以在个人需要在家进行医院检查或自我隔离的情况下为他们提供建议。
This can limit unnecessary access to healthcare facilities and reduce the workload on the medical staff who are hardly finding time for themselves and their families since the pandemic has started. Using Al for quick diagnosis can significantly help in flattening the curve in any country. This can efficiently prevent the disease from spreading and provides a key for high-throughput diagnosis. Applying good prevention techniques can help prevent the healthcare system to avoid flooding the hospitals.
这可能会限制不必要的医疗保健设施使用，并减少自大流行开始以来几乎没有时间为自己和家人服务的医务人员的工作量。 使用Al进行快速诊断可以显着帮助在任何国家/地区展平曲线。 这可以有效地防止疾病传播，并为高通量诊断提供了关键。 应用良好的预防技术可以帮助防止医疗保健系统避免洪水泛滥。
2.阿里巴巴针对COVID-19的CT图像分析： (2. ALIBABA’S CT IMAGE ANALYTICS FOR COVID-19:)
Problem- As a part of the diagnosis process, doctor’s take around 2000 Computed Tomography (CT) scans of a patient in the entire hospitalization process. Manually reading and comparing these scans requires a huge amount of manual effort. It takes around 15 minutes for an experienced doctor to analyze these scans. As the number of patients has proliferated considering the situation of how the virus is spreading there has been a scarcity in the number of doctors and experienced professionals. Medical professionals are made to work overnight and the individuals assessing these scans are often exhausted and operating under exorbitant workloads where human error could easily become an issue.
问题-作为诊断过程的一部分，医生在整个住院过程中对患者进行了约2000次计算机断层扫描(CT)扫描。 手动读取和比较这些扫描需要大量的人工。 经验丰富的医生大约需要15分钟来分析这些扫描结果。 考虑到病毒传播的情况，随着患者数量的激增，医生和经验丰富的专业人员的数量已经不足。 医疗专业人员需要整夜工作，而评估这些扫描的人员通常筋疲力尽，并且在过高的工作量下工作，而人为错误很容易成为问题。
Solution- Hence, there is a need for accurate CT scan analysis system. To expedite this scrutinizing process, Alibaba had come up with a CT IMAGE ANALYTICS system which can help with the significantly improving the testing accuracy and detection efficiency rates. This system has been able to assist in identifying the coronavirus pneumonia CT scans with about 96% accuracy. This identifies the scans within 20 seconds, which is 60 times faster than manual human detection methods. This was developed by Alibaba’s research and innovation team called DAMO Academy.
解决方案-因此，需要精确的CT扫描分析系统。 为了加快审查过程，阿里巴巴提出了CT IMAGE ANALYTICS系统，该系统可以帮助显着提高测试准确性和检测效率。 该系统已能够以约96％的准确度帮助鉴定冠状病毒性肺炎CT扫描。 这可以在20秒内识别扫描，这比手动人工检测方法快60倍。 这是由阿里巴巴的研究和创新团队DAMO Academy开发的。
This Al system has been trained with a dataset of more than 5000 positive coronavirus cases. It uses deep learning algorithms to study the patterns of infection and can be used for two things effectively: one is to track responses in confirmed cases, and the other is to provide diagnoses for the suspected cases. The pre-eminence of this algorithm is that it can find the difference between common pneumonia and COVID- 19 pneumonia. Once this system identifies a CT scan as positive coronavirus case, it can also help further by providing analytics like the output of the lesion marks and affected lung volume ratio. This helps the doctors to productively track and look at the development of treatment of the Covid-19 patients. More than 160 public institutions in China are using this technology and has already analyzed more than 240,000 CT scan.
该Al系统已经用超过5000个阳性冠状病毒病例的数据集进行了训练。 它使用深度学习算法来研究感染的模式，并且可以有效地用于两件事： 一是跟踪已确诊病例的React，另一是为可疑病例提供诊断 。 该算法的主要优势在于它可以发现普通肺炎和COVID-19肺炎之间的差异。 一旦该系统将CT扫描识别为冠状病毒阳性病例，它还可以通过提供诸如病变标记的输出和受影响的肺体积比之类的分析来进一步帮助。 这有助于医生有效地跟踪和观察Covid-19患者的治疗进展。 中国有160多个公共机构正在使用该技术，并且已经分析了超过24万个CT扫描。
用于智能机器人和无人机监控的AI (AI for Monitoring by Intelligent Robots and Drones:)
In many countries around the world, wearing a mask in public has been made mandatory by the governments. Countries like China and India impose heavy fines on people who are roaming outside without wearing a mask. To make this process easier, some cities and governments have accelerated the public deployment of Al powered drones. They can effectively track individuals who are not wearing protective masks which is a scientifically proven remedy to deaden the spread of coronavirus. Also, China is using is Al-based multisensory technology which is deployed in public places such as airports. This system can screen up to 200 people per minute and can detect their temperature within a range of 0.5 degrees Celsius. This allows the airport authorities to quickly isolate a person who is suspected of having a fever.
在世界许多国家/地区，政府已强制在公共场所戴口罩。 中国和印度等国家/地区对未戴口罩的在国外漫游的人处以高额罚款。 为了使此过程更容易，一些城市和政府已加快了Al动力无人机的公共部署。 他们可以有效地跟踪未佩戴防护口罩的个人，这是科学证明可以缓解冠状病毒传播的补救措施。 此外，中国正在使用基于Al的多传感器技术，该技术已部署在机场等公共场所。 该系统每分钟最多可以筛查200个人，并且可以检测他们在0.5摄氏度范围内的温度。 这使机场当局可以Swift隔离怀疑发烧的人。
Problem- I am sure many of you might have heard the news about many healthcare workers succumbed to coronavirus. This is due to the increased exposure of the workers to the corona affected patients in this hospital.
问题-我相信你们中许多人可能听说过许多医护人员死于冠状病毒的消息。 这是由于工人在该医院中接触受电晕影响的患者的风险增加。
Solution- This can be solved only when we decrease the number of interactions between doctors and patients. Autonomous vehicles are also playing a crucial role in distributing food and medicines to patients in hospitals. CloudMinds deployed more than 100 robots in Chinese hospitals, hotels and schools.
解决方案-仅当我们减少医生与患者之间的互动次数时，才能解决此问题。 自动驾驶车辆在向医院患者分发食品和药品方面也起着至关重要的作用。 CloudMinds在中国的医院，酒店和学校中部署了100多个机器人。
These Robots embrace a humanoid called Cloud Ginger (XR-1) and the Smart Transportation Robot, can carry food and medicines to patients from health-care providers without person-to-person contact. Since these are Al powered robots, the bored and isolated patients have found to enjoy his company as they offered useful information and much-needed interaction and entertainment. Its ability to provide real patient vital signals remotely without direct contact has been found extensively fruitful for the doctors.
这些机器人包含名为Cloud Ginger(XR-1)的类人动物和智能交通运输机器人，可以将医疗食品提供者的食物和药品运送给患者，而无需进行人与人之间的接触。 由于这些都是Al动力机器人，无聊和孤立的患者发现他喜欢他的公司，因为他们提供了有用的信息以及急需的交互和娱乐。 已经发现，它无需医生直接接触就可以远程提供真实的患者生命信号的能力，对医生而言非常有用。
使用AI预防下一次灾难： (Using AI to prevent the next plague:)
“ Prevention is better than cure ”
“ 预防胜于治疗 ”
Coronavirus is not the first pandemic the world has seen. There have been many pandemics the world has faced and the most devastating one is the 1918’s Spanish flu which has affected almost around 500 million people and had caused around 5 to 7 million deaths. But, as the technology is growing, we can reduce the number of deaths and infections using the latest scientific advancements and research happening around the world and ostensibly, using AI-powered systems has proven to be a lifesaver. We can follow the below-mentioned series of steps to thwart the next possible plague.
冠状病毒并不是世界上第一个大流行。 世界上已经遇到了许多流行病，其中最具破坏性的是1918年的西班牙流感 ， 该病已影响将近5亿人，并造成约5至700万人死亡。 但是，随着技术的发展，我们可以利用世界各地发生的最新科学进展和研究减少表面上的死亡和感染数量，事实证明，使用人工智能技术的系统已被证明可以挽救生命。 我们可以按照下面提到的一系列步骤来阻止下一次灾难。
Step 1 — Predicting :
第1步-预测：
Predicting a pandemic before-hand and preparing for it is the best preventive step we can take. Bluedot and Metabiota are capitalizing the power of NLP algorithms to scan and monitor several news outlets around the world. The flag and mention high priority diseases such as coronavirus as pandemic possible outbreaks and flag endemic ones such as HIV or tuberculosis. So, predicting a next possible occurrence of plague will help the government and countries across the world to tackle the drastic situations with building new hospitals, imposing lockdowns and travel restrictions, increasing the availability of medical staff and hospital beds etc, will help to terminate the spread of the plague to all the countries around the world.
事先预测大流行并做好准备是我们可以采取的最佳预防措施。 Bluedot和Metabiota正在利用NLP算法的功能来扫描和监视世界各地的多个新闻媒体。 船旗国并提到高度流行的疾病，例如冠状病毒，可能导致大流行；而船旗国特有的疾病，例如艾滋病毒或结核病。 因此，预测下一次可能发生的瘟疫将帮助世界各国政府和国家通过建立新医院，施加封锁和旅行限制，增加医护人员和病床的可利用性等方式来解决严峻的形势，将有助于终止传染病。瘟疫蔓延到世界各地。
Step 2 — Finding Cure :
第2步-找到治愈方法：
Al can help researchers scour thousands of research papers and documents around the world and can unwrap possible cures and vaccine which is a challenging task for humans. Traditional vaccines can take up to ten years to develop and we cannot afford to wait for that much time and look at the people and economy to become moribund. A key step in creating a cure is understanding the underlying structure of proteins and makeup antibody proteins which researchers can use for efficient drug discovery while reducing the costs and time associated with these findings. Google Deepmind’s AlphaFold is using Al for the scientific discovery of predicting possible protein structure of the antibodies with a hope of making a real-world impact.
Al可以帮助研究人员搜寻世界各地成千上万的研究论文和文件，并可以包装可能的治疗方法和疫苗，这对人类而言是一项艰巨的任务。 传统疫苗可能需要长达十年的时间才能开发出来，我们不能等待那么多时间来关注人民和经济的mo折。 建立治疗方法的关键步骤是了解蛋白质和补充抗体蛋白质的基本结构，研究人员可利用这些结构来进行有效的药物发现，同时减少与这些发现相关的成本和时间。 Google Deepmind的AlphaFold正在使用Al进行科学发现，以预测抗体的可能蛋白质结构，希望对现实产生影响。
结论： (Conclusion:)
While there have been claims around the world about data privacy, for an Al system to make accurate predictions, it should be trained with different types and classes of data. After all, an Al system was the first one to identify that the COVID-19 has potential pandemic forming capabilities. This demonstrates that Al along with the collaboration of the best scientific minds around the world will be most efficacious in controlling, identifying and monitoring a COVID-19 and any other disease in the future. We hope that AI will be useful for disease research, diagnosis, monitoring the patients and ultimately improve the quality of life for millions of patients around the world.
尽管世界各地都有关于数据隐私的主张，但要使Al系统做出准确的预测，应使用不同类型和类别的数据对其进行培训。 毕竟，铝制系统是第一个识别出COVID-19具有潜在大流行形成能力的系统。 这表明，在未来，铝与世界各地最先进的科学头脑的合作将最有效地控制，识别和监测COVID-19和任何其他疾病。 我们希望AI将对疾病的研究，诊断，监测患者有用，并最终改善全球数百万患者的生活质量。
翻译自: https://towardsdatascience.com/how-ai-and-ml-are-helping-to-combat-covid-19-fe8ecbb88b7e
php-ai/php-ml
401
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
