Following are five simple use cases that represent a fraction of the potential of artificial intelligence. While adoption in some areas has been faster for multiple reasons others represent a potential to be exploited.
Support for children with autism
Autistic children and their families have to deal with many problems, including communication, which plays a crucial role in everyday life. A community of parents, researchers and specialized educators around the world have recently turned to artificial intelligence for new solutions to support children with autism.
In the first case, the Natural Language Processing (NLP) capabilities were used to suggest the right answer to the autistic child being asked a question. The child is equipped with a light equipment that contains a microphone, an earpiece and a communication module. The microphone picks up the question asked by the interlocutor (for example: What is your name?). The question is then transmitted by the communication module to the NLP module which analyzes the question and finds the corresponding answer to route it into the child's earpiece. This method of voice assistance can accompany the autistic child in real time in an uncontrolled environment.
In a second case, a group of occupational therapists and engineers developed a bracelet that allows taking physiological measurements (pulse, blood pressure, sweating, movements). This haptic device makes it possible to detect the patterns that announce problematic stress episodes to potentially avoid them. An avatar using NLP on a digital tablet combined with neurotransmitters on the haptic device are used to communicate with the child and bring him back to a calmer state.
Seeking information
In 2018, more than ever, we have access to a multitude of public and private documents that abound with useful information for our work. For most of us, getting well informed has become second nature. In some areas, however, the amount of documentation is such that it becomes impossible to absorb everything.
Oncology research alone produces 10,000 documents per month. In the legal field, some litigators spend 50% of their time reading documents to find laws, regulations, judgments or rulings. And these are just two examples among the myriad of domains that require analyzing documents. This part of the work as much at the doctor as at the lawyer is not a pleasant task but rather a necessary evil and since in both cases it is not possible to read all the documents, it was necessary to find a way Automate reading and filter information to extract relevant passages that will be useful depending on the context.
This is where artificial intelligence comes into play with its Natural Language Processing (NLP) capabilities. An NLP system trained for a specific area (scientific, legal, etc.) will be able to read thousands of documents in a few minutes and will be able to select the relevant sections to bring them to the attention of the user. This is a perfect example of the use of the cognitive capacity of language by artificial intelligence in collaboration with humans. You can see the capabilities of such a system on
rossintelligence.com
Agricultural industry
The Industry 4.0 transformation wave also involves the use of artificial intelligence.
For example, in the agricultural field, several technologies are pooled to make spectacular advances in the
control of production. Take the case of this company that has developed a solution to water lettuce plants using
a method that significantly reduces the use of chemicals. During periods of watering of insecticides and
fertilizers, the solution equipped with cameras makes it possible to analyze the state of each plan of
lettuce individually in real time to apply with the aid of nozzles of high precision the quantity of
insecticide or fertilizer required. Each lettuce is listed in the system to monitor the progress of each
plant until harvesting. The AI algorithm that makes the watering decisions based on the appearance of the
plant was trained from thousands of photos of lettuce plants that had various deficiencies.
Medical imaging
One of the cognitive functions of Artificial Intelligence (AI) is its ability to recognize
the content of a photo, an X-ray, a movie. Rapid developments in the field of deep learning neural networks
have led to solutions that will change the way we work in the years to come. The field of health is particularly
conducive to the use of these solutions.
For example, AI systems have been trained from thousands of images to recognize dermatological abnormalities that
could be cancerous. During an examination, the family doctor who detects a suspicious lesion on the patient's skin
uses a smart phone to which he adds a removable magnifying glass to take pictures of the lesion that he or she sends
to an AI system for analysis. After a few seconds the system returns a result with a percentage of probability that
the lesion is cancerous. In the case where the result is positive, some clinics have agreements with dermatologists
to send the photo of the lesion so that the patient passes in priority if the diagnosis is confirmed. This new procedure
avoids months of waiting and anxiety for patients who have been referred to dermatologists.
Conversational agents
A conversational agent, also known as 'Chatbot', is a solution that is often deployed in call
centers, on shopping sites or as a concierge service to inform customers in hotels.
The conversational agent can be developed with varying degrees of sophistication from simple question-based
answers to corporate information access for more complex questions. These agents may also be equipped with
tools to analyze the words used by the interlocutor in order to profile his state of mind and adapt the exchange
to establish better communication. For example, if the conversational agent determines that the caller is showing
signs of impatience, he or she will be able to route the call to a real person who has been previously informed
about the context of the exchange and will be able to take over the call in a more harmonious way. Conversations
with the conversational agent can be done in text format or using speech. Some specialized conversational agents
developed to make appointments using the voice have reached a level of language proficiency quite surprising.
A striking example is available on this site :
site:gstatic.com