Artificial intelligence is a word that is currently rolling off everyone’s tongue. Although it has been around for over sixty years, AI is the expression of the hour. Sometimes using the terms “machine learning”, “deep learning” or “algorithms” as synonyms, the media hype surrounding AI would have every reason to cause skepticism. The fact remains that these techniques are extremely important in today’s world. The level of power reached by computers has resulted in a definite arrival of artificial intelligence.
We are not saying that AI is about to put the entire planet on unemployment or that it has reached the point of singularity; that moment when AI will become a superintelligence that will by far surpass that of humans and will compromise their power on Earth. Rather we would like to say that, as mentioned in our previous article, reflecting on the CRM technological trends of 2017 the AI will continue to integrate itself exponentially in all spheres of our daily lives. Since AI will become fundamental to our functioning, it is crucial for entrepreneurs to grasp its challenges now.
These four crucial definitions will guide you to a better understanding of AI, each definition will be followed with examples of present-day applications.
This concept is the largest one. Artificial intelligence is defined as any system capable of accomplishing tasks in a manner that is perceived as intelligent. Thus, any device that simulates aptitudes deriving from human intelligence such as abstraction, creativity, deduction, problem resolution, decision making or the ability to learn are all linked to AI.
As a global concept, AI is a branch regrouping various forms of artificial intelligences. The AI branching tree below demonstrates this. Avoid however looking at these ramifications as parallel sub-branches. Rather, think of them as climbing plants: often codependent and knotted.
Machine learning is often what comes to mind when it comes to AI. As a matter of fact, these two terms often replace each other as two socks of one pair. This AI application is often described as the capacity of imitating human learning aptitudes. Instead of reporting to a program that tells the system how to react in any imaginable circumstance, the system is literally auto-teaching itself.
How do we create such good students? By providing the systems with data from which they can learn and act autonomously according to the gathered information. This process is made possible using statistical analysis and repetition learning.
For example, the more you log into your Facebook account, the more the algorithms can customize the content in your newsfeed. This is also the case with sites such as Amazon or Netflix where they provide you with suggestions on what books to buy or what series to follow. These sites gather information regarding your usage and compare them to other customers who liked the same products. The suggestions become more adapted to your interests and satisfaction as your usage increases.
This is another popular term often used as a synonym of AI. This concept, however, is a recent one since it stems from the advancement of the power gained by computers. This technique is still developing and is being greatly invested in by key players such as Google, Apple and Microsoft. As with the term machine learning, when you hear the expression AI it is often related to deep learning. Nothing surprising since the two are naturally linked.
Deep learning comes from a machine learning implementation technique. It usually works by means of algorithms that try to reproduce the structure of the human brain called neural networks. The functioning of neural networks is inspired from our own capacity to process information in a non-linear association mode.
One example of a deep learning application is DeepL, a machine translation site which was launched last August. You will appreciate its preciseness when translating texts or even simple phrases. As a matter of fact, it has become the preferred website for translators since contrary to Google Translate, DeepL seems to grasp the context of the words. The secret? It makes comparisons with similar contexts where word sequences have already been used. DeepL has access to Linguee (built by the same company) with a rich repertoire of royalty-free texts available in several languages. DeepL is then able to process these results using an ultra-powerful supercomputer located in Iceland.
We decided to test Google and DeepL translators to compare their performances. To stay within the theme of AI, we chose an excerpt from the novel 2001: A Space Odyssey. The results are significant:
But nothing is more affirming than a test: try it for yourself.
To summarize, a chatbot (also be called a conversational agent) is a robot capable of imitating a dialogue between humans either through voice or text channels. The functioning of this interlocutor includes both machine learning and natural language processing (NLP).
There is a good chance that you already have called upon a device with the name “Siri”,”Alexa” or “Hey Google”. Vocal assistants have gained in popularity and have hit record highs in sales over the holiday season. Comparably, the popularity of chatbots has spread at an unprecedented speed on websites of companies since they are very appreciated for their top of the line customer service. You might have already conversed with one of those chatbots unknowingly.
Finally, a more lucrative example would be Replika a chatbot used for leisure. This free application offers a virtual friend ready to listen at any time of the day. The more time you spend conversing with Replika, and evaluating either its positive or negative responses, the more this chatbot will get to know your taste and trace your personality even imitating the way you express yourself.
These are the main differences between AI, machine learning, deep learning, and chatbot. Of course, this is a non-exhaustive glossary. But these four concepts serve as a good basis to better understand the field of artificial intelligence which will permanently transform business processes in the years to come. This will be the topic of our next article.
For any questions, do not hesitate to contact one of our experts who will be able to give you personalized advice.
Create a customer engagement environment for your customers with Dynamics 365 (formerly known as Microsoft Dynamics CRM)
Your customers expect you to deliver a personalized and enriched customer experience . XRM Vision will manage this interaction with you by putting at your disposal the best technology to help you with customer retention, improving your productivity and managing your growth. Learn how.
Based on the Capability Maturity Model, our methodology will help you assess the CRM maturity level of your business processes by reading our e-book and completing our online questionnaire.