Unpredictability to Predictability: AI Set to Transform Logistics, Medical, and Geographical Sectors

By Pratik Kirve

Artificial intelligence (AI) is nothing short of wonder that amazed the every sector across the world with its potential. The unlimited application scope and the unbounded potential of AI can bring transformation and revolutionize the processes in each sector including healthcare, manufacturing, environment, and others. It can accelerate the research and development activities and predict unimaginable things. From predicting the appropriate routes for logistic operations to predicting sea ice conditions in Arctic, the applications of AI have been increasing day by day. An AI platform has been developed by a tech firm that operates like a virtual assistant and offers necessary insights and advice to drivers regarding routes. Moreover, researchers developed a new system that can predict the Arctic sea ice conditions to determine the potential risks. 

The healthcare sector would reap great benefits of advancements in AI. From determining the metastatic risks in patients suffering from skin cancer to identifying the diseased cells, it will prove to be vital in devising the diagnosis and treatment course. The AI technology would transform almost every sector in the coming years. According to the report published by Allied Market Research, the global artificial intelligence market is estimated to reach $169.41 billion in 2025. Following are some of the activities taking place across the world. 

The innovations in the world of AI are going beyond the scope of imagination. Tech companies have been utilizing the potential of AI in different applications. In the logistics space, Simpliroute, a tech firm from Chile, developed an AI based virtual assistant ADA. This virtual assistant would help drivers in determining the routes, the selection of vehicles, and the potential risks in the routes taken through the historic data. The company integrated the AI platform for logistics operations. Drivers and fleet operators can avail a lot of benefits through the data and insights provided by the assistant ADA. The company already launched an app and drivers have been working with ADA to gain alerts and advice on different routes. Moreover, it suggests the changes to be made in the fleet size to fleet operators. The company has been working on launching the updated version of the app and improving the integration with machine learning algorithms. 

The application scope of AI is widening and it can be used for forecasting the geographical conditions. Estimating Arctic sea ice conditions is possible with new AI system developed by the team of researchers from British Antarctic Survey (BAS) and The Alan Turing Institute. The new AI system known as IceNet can offer accurate forecasts of the upcoming season regarding sea ice conditions of Arctic. Sea ice is the frozen sea water present at South and North poles and predicting the atmospheric conditions had been difficult for researchers. With the development of IceNet, it is possible to predict the presence of sea ice two months prior to its appearance. The accuracy offered by the system is nearly 95%. 

Tom Andersson, the Data Scientist at the BAS AI Lab and the lead author of the study, highlighted that the new system can bridge the gap in estimating sea ice to accelerate the sustainability efforts. This system gives predictions nearly thousands of times faster in comparison to traditional methods. Deep learning algorithms have been used to drawing predictions. From the simulation data on sea ice changes in thousands of years and the observational data of past decades, the system can predict the sea ice conditions of the near future. The AI system is capable of giving an early warning regarding the risks related to rapid loss in sea ice.

The healthcare sector is expected to be benefitted greatly by AI tools and systems. Researchers have been developing new methods of utilizing AI for identifying diseased cells. Researchers from Helmholtz Zentrum München and Munich University of Applied Sciences (TUM) developed an algorithm known as ‘Single-Cell Architecture Surgery (scArches). In this algorithm, the process known as mapping is used. In this process, transfer learning is used for comparing new data sets about the single-cell genomics with the available references. The reference atlas of healthy cells is compared with the cells of sick patients to determine the diseased cells. In addition, the algorithm maintains the anonymity and privacy of patients.  

The scArches algorithm has been utilized for determining the diseased cells in the patients infected with the Covid-19. The lung cells of Covid-19 patients are compared with the lung cells of healthy individuals. In this manner, researchers were able to determine affected cells in both mild and severe cases of Covid-19. The biological variations in patients had no impact on mapping. Highlighting the significance of scArches, researchers said that they aim to simplify and formalize the search process and identification for devising a better treatment course.

AI has been making strides in the healthcare sector by proving its effectiveness in cancer and other diseases. Researchers from the UT Southwestern Medical Center developed a method to predict the metastatic risks in skin cancers using AI. They determined the efficacy of AI tools in revolutionizing the study for cancer. A general framework has been developed by researchers to determine the progression of cancer toward metastatic states through collection of tissues samples and predicting mechanisms. 

The team of researchers used AI tools for identifying the difference between images of melanoma cells that have high and low metastatic potential. Then they reverse-engineered the findings from the AI tools to determine the distinguishing features. Tumor samples from seven patients of skin cancer were collected along with information on the progression of disease. After filming the video of nearly 12,000 cells in petri dishes, around 1,700,000 raw images were generated. Then AI tools were used for extracting 56 abstract numerical features from these raw images. Among these features, researchers were successful in finding a single abstract numerical feature that differentiated between the low and high potential of metastatis. Then artificial images were created by manipulating that feature and characteristics of metastatis were exaggerated. They found that cells with the high metastatic potential created increased light scattering and more pseudopodia extensions in comparison to cells with the low metastatic potential. This experiment highlighted the usage of AI tools in determining the metastatic risks in the cases of skin cancer. 

About the Author

Pratik Kirve is writer, blogger, and sport enthusiast. He holds a bachelor degree in Electronics and Telecommunication Engineering and currently working as a Team Lead – Content Writing at Allied Market Research. He has avid interest in writing news articles across different verticals. When he is not following updates and trends, he spends his time reading, writing poetry, and playing football. He can be reached at [email protected]