Four Takeaway of Those of You Who Want To Enter AI Industry
Artificial intelligence has experienced some gigantic progress during the past several years worldwide.
But it seems disappointing. And today’s greatest example is artificial intelligence, surrounded by hype and floods of finance. Beyond science and engineering, there are hobbies which muddy and hide the true job of individuals. Although the field is very beautiful. It grows and develops unbelievably quickly and promises to maintain that speed for some time. In the next several years there will be more AI-related employment. And, let’s be honest, it’s far more thrilling and fascinating than most of the other areas — we’re trying to figure out, after all.
AI allows computers to learn and act in lieu of humans or to complement people’s job. We have already seen broad usage of AI in our everyday lives, for example when companies like Netflix and Amazon are presenting us with choices based on our purchasing habits or when chat bots answer our questions. AI is utilized for piloting aircraft and even streamlining our traffic signals.
Therefore, many of you may benefit from knowing what to anticipate. Here are four things that I wish I understood before I started AI. I was psychologically prepared for what I was about to encounter and had a more fine, deeper view of the field today and the future. Enjoy!
“The AI effect” tries to redefine AI as “anything not yet done.”
Well this is an overall trend in this area. AI is a term that sells — yet people grow weary of it increasingly, and rightfully so. However, most AI goods and services catch the attention of investors because they want to hear the terms “artificial intelligence.”
It offers a great deal. And in itself, that’s not terrible. But when more businesses claim AI answers than do anything to push things forward — or apply the technology to real-life problems — then people mistrust.
London venture capital company MMC concluded in 2019 that 40% of European “AI companies” did not utilize AI. “We could not uncover any indication of AI in 40% of instances,” stated MMC Research Leader David Kelnar. “[C]companies that people believe are probably not AI businesses.”
Beware of anything in its name that contains AI. It’s not everything that purports to be AI. And it simply damages the confidence and credibility of those that develop genuine AI.
Deep-learning AI’s are so easy to fool
I am not saying that DL is not solving the tasks: DL is impressive
Most of us learned about AI together with deep learning and machine learning. We are likely to associate these ideas with each other. AI is, however, far older than profound learning. The phrase artificial intelligence was created by John McCarthy in 1956, while the first time someone mentioned “deep learning” related to AI was in 1986.
As it is a scientific field, AI has a special characteristic. This is not constant. It is not constant. Whatever comes under the “AI system” category now may not be in a few decades’ time. It appears that only what we don’t comprehend is AI. Deep learning systems tick all the criteria for impressiveness, flair, complexity and futurism. But not everything in AI is like that.
In the mid-1990s, many sectors started integrating rudimentary machine learning applications without us recognizing them following the second AI winter. The mainstream media ignored it mainly because it lacked the fancy we usually attribute to AI. We are impressed and amazed by deep learning systems like AlphaZero, GPT-3 or Codex. They remind us of science fiction and a possible future in which machines coexist. This is the kind of vision that we collectively assign to AI. Linear regression is mostly not AI.
Deep learning is undeniably the only effective paradigm since AI was created. However, we should constantly bear in mind that AI is wider than profound learning. Not just as a technology or science, but also as a philosophical lens to comprehend the meaning of the human being and the mysteries of understanding. Equating both does the field a disservice.
The hybrid method is a potential new paradigm that seeks to use the best of profound learning and look at previous ideas to avoid the bottlenecks of today. Nobody knows what AI is going to look like in 50 years, but it is just irresponsible to ignore ancient concepts which have strong theoretical foundations and which may help build a basis for the next major breakthrough.
This is a basic Venn diagram that provides a view of the AI-deep learning relationship:
THE TOP MYTHS ABOUT ADVANCED AI
AI is not as sophisticated as the mass media representation
A fascinating debate rages on concerning the future of AI and what it means for mankind. The future effect of AI on the labor market; if/when human-level AI will be created; if this will lead to an intelligence explosion; and whether we should embrace or dread this.
But there are numerous instances of dull pseudo-controversies created by miscommunication and misunderstanding. Let’s dispel some prevalent misconceptions so we can concentrate on the fascinating debates and open issues, not the misunderstandings.
There is an ancient tendency in the area of overpromising and under-performance. It was so prevalent that funding ups and downs had a name: AI summers and winters.
Now, it’s not fair to blame the media for everything. The founding fathers were confident that in a few of decades they could create a general artificial intelligence. In 1967, before the first AI winter, Marvin Minsky, perhaps the most renowned of them, stated that “within a decade, the challenge of producing ‘artificial intelligence.’
Then it didn’t happen. It didn’t happen yet. The most likely explanation is that we all misjudge the scale of the task. And there’s also a reason for it; we don’t comprehend the intellect and the brain we don’t understand. How can the difficulty of a task be properly estimated if we don’t grasp the rules?
AI specialists — some more than others — have learnt from it and have chosen to keep their promises to a low standard. Their objective is now not to generate interest in the area, but it is feasible to fulfil what they promise. Nevertheless, there is another significant player whose objective is not very much in line with that. The mainstream media don’t worry much about AI’s lack of credibility. They worry about the headlines and are kept in interesting articles by their readers.
Science nourishes the truth. Media attentiveness feeds.
Let’s see a few news about current advances in AI.
- “The whole piece was written by a robot. Are you still frightened, human?” What happened: eight pieces were written by GPT-3. Then, The Guardian “selected the finest bits of each one” and merged them to create the final piece, perhaps to make greater impact.
Also, if want to know more about GPT-3 here you can have a check on my article here “My First Hands on Experience with GPT-3!”
- “Robots Can Read Better than Humans Now, Endangering Millions of Jobs.” What happened: an AI system overcame people with a basic reading assignment. Even the designers of the exercise stated it was not intended to assess the human meaning of the term for reading.
Or AI anthropomorphic headlines.
- Verizon: “How AI knows emotion.”
- Engadget: “MIT develops an AI that instinctively knows the rules of physics.”
- KDnuggets: “Construction of an AI that knows the world through video.”
AI does not comprehend, think or feel. It does not understand. It accomplishes it differently than humans do, whatever it does.
Technically speaking, the titles are not incorrect, but they use ambiguity and ignorance to capture opinions. Or, worse, the editors deliberately conceal some of the data to decorate the moral blameworthy piece. We have left this confused and our ideas mismatched with reality after reading it.
Finally, in recent times, I’ve noticed a small change from these behaviors. People were weary of flagrant clickbait and news agencies were noticed. Hopefully this destructive tendency will dry up soon.
Technology And Science Come Together In Artificial Intelligence
Artificial intelligence may be understood from two viewpoints.
- It was a field of engineering that concentrated on creating useful applications that enhance the well-being of people. In Google, Amazon, Netflix, Spotify, we have numerous instances. All these businesses rely heavily on their AI technologies that we use every day.
- AI may be seen as an area of science. The founding fathers of AI did not take the current environment into account when they officially established the field. They desired a universal artificial intelligence to be built; an AI as clever and competent in every respect as humans. They believed it was an achievable objective and so every effort was directed in this direction.
The short-term objective of creating helpful applications for individuals in their daily lives and of improving our collective knowledge of intelligence and cognition is fully compatible. However, we need to remember that the two aspects of AI are distinct, and as such they grow and take place at different levels.
Despite the danger that AI could cause future job losses, AI also has the ability to transform our world for the better. AI can minimize human distortions and errors in data analysis, improve efficiency, relieve employees from repetitive activities and more at company level. But it can bring much more benefit worldwide.
Some initiatives concentrate on the application of science to create beneficial technology. Other attempts are aimed towards furthering the discipline without paying attention to the immediate applicability of the findings. One is transported by profit. Curiosity moves the other.
Everything you select is good, just don’t become confused. And most all, don’t assume they go together naturally. Learn to integrate both perspectives using the multi-dimensional AI idea. You will eliminate all the uncertainty about where and where the field is today.
Takeaways
- Not every “AI” is AI. We unconsciously associate the term “AI” with deep learning algorithms and neural networks. Some companies use marketing to attract investors, customers, or employees. Watch it.
- AI isn’t simply deep learning. Deep learning is the most successful AI paradigm today, but it hasn’t always been that way, and it may not be in the future. You won’t be lost if deep learning returns to previous ideas.
- AI is not as sophisticated as the portrayal of the mainstream media. Historically, AI specialists tended to overpromise what they could do. The mass media now frequently follow the same pattern. Headlines overstate what is done in such a manner that readers are misled. From the outside, we can only know from what the media tells us about what is occurring in AI. Try going to the major sources and judging yourself.
- AI is a technical and scientific discipline. This ambiguity contributes to some of the confusion. We utilize AI in everyday applications, but it is also used in scientific study to uncover some of humanity’s biggest mysteries: intellect and the brain. Combine the two to get a broader understanding of AI.
If you found the article useful, please feel free to hit ‘CLAP’ button — multiple times. Go nuts and do comment! Leave feedback! :)
