The Imminent Arrival of the Super Chatbots

“Most tech giants are investing heavily in applications and research, hoping to stay ahead of the curve of what many believe to be an inevitable AI-led paradigm shift” [1]

The New Breed of Chatbots

So what is heading the cutting-edge upgrade of chatbots and their interactive conversations? The answer to this is: self-governing navigation and computer vision. Now, due to revolutionary machine learning methods, high tech hardware and data availability, this has meant that their usability has been completely transformed. Today, advances in artificial learning are modifying issues that were until recently, deemed impossible to resolve. And they are gradually empowering our daily lives in all kinds of spheres [1] including media.

AI is Rapidly Heading Up the Popularity Charts

How can we define a bot's intelligence? “A chatbot is intelligent when it becomes aware of users' needs. Its intelligence is what gives the chatbot the ability to handle any scenario of a conversation with ease” [2]. In 2017, the market was overrun by a new chatbot generation - one that promised us a world with meaningful interactive conversations. “Helping turn this promise into reality is a combination of better user interfaces, the omnipresence of smartphones, and new, state of the art machine learning techniques” [1]. Deep Learning is a huge driver boosting this tide of unique AI applications - a sphere of machine learning which not so long ago, revolutionized areas including NLP (natural language processing), and computer vision, even though deep learning has been on the scene for fifty years or so [1].

Can Deep Learning Alone, Help Chatbots' Challenges?

The simple answer to this is no! In order to conduct and maintain a meaningful conversation with a human, the humanoid must be programmed to comprehend the context of what is being said, and this involves having abilities such as: logical reasoning, being aware of the consumers' intent and preferences, and being able to disambiguate a language's subtle deviations which can have radically different meanings [1].

Further, in the realms of super powerful chatbots, is is essential for the latter to be able to reply to intricate questions, not just using context, but also information that goes above and beyond the dialog restrictions. In order to reply to users' questions efficiently, until now, the chatbot has been completely reliant on previously shared conversational data between the user and itself. Further, a company's strategy and the users' intent also play a role in the type of reply the humanoid will come up with. Yet, if the bot needs to evolve to do more than just answer easy, one dimension questions: “it must blend the most prominent techniques emerging from the field of deep learning with solid statistics, linguistics, other machine learning techniques, and more structured classical techniques, such as semantic parsing and program induction” [1].

Data is King

Data is the first essential stage for constructing an intelligent conversation structure. In order to achieve its full potential, deep learning requires massive quantities of high quality data. Yet although there is a perpetual flow of eternal data streams, when it comes to machine learning algorithms, the majority of it is just too raw to be used as is. The sub-field of machine learning - unsupervised learning, is a potential solution to this dilemma. This is dedicated to extracting raw data information, and is something that does not require human input. In a nutshell, it transforms the raw data into a less complex form, thereby enabling an easier way to discover patterns [1].

Obstacles

One of the great benefits of deep learning, is that it offers a state-of-the-art performance, by providing cutting-edge ways to build. Once these techniques are optimised, it removes the necessity for a great deal of expensive and high quality labeled data – something that has been mandatory to ensure that chatbots perform well. There are however, a few obstacles in the way. - Deep learning is reliant on collecting a huge dataset which is extremely specific, followed by training a network [1].

While this is the basis for massively powerful systems, it does not have the flexibility to generate a natural conversation. Further, another hurdle in the search for bots who are really human-like, is empowering them with the means to: “reason with an internal model of the world. The level of automatic and broad reasoning that we humans have still eludes AI researchers, and is one of the last frontiers in the way of truly intelligent and autonomous... conversational bots” [1].

Ultimately, a conversational engine needs to be more precise. It is not enough to just utilise NLP to determine an arrangement of relevant data. The chatbot must have the means to analyse the input, “break it down, and present a response to the user that is not only clear and concise, but highly relevant to their taste — rinse and repeat” [1].

The Future of AI-Driven Conversations

“Any type of AI should one day be able to recognize how you’re feeling, and then use that information to respond meaningfully, mirroring a human’s emotional state the way another human would” [3]

At this moment in time, the revolution of AI-driven conversation is in its infancy. To that end, it is very likely that high tech innovations will surmount the current issues [1]. Users such as those who connect with Microsoft's personal digital agent, Contana, and consumers who get their entertainment from platforms such as OONA Mobile TV, the innovative AVOD OTT tech platform with premium options, which pioneered the patented AI Genie in the App, Siskabot, will before too long, see the day when they will be enjoying long intricate interactions with their humanoid counterparts. The bots will not just thoroughly understand the consumers and what they need, they will also do as OONA already does – know what their preferences are, and adapt their experience accordingly.

So What Makes OONA so Different?

OONA, the World 1st Entertainment & Messaging Platform that gives users a cool way of choosing, sharing & editing their fab entertainment, messages & ads, is delivering a high tech customer experience (CX), to engage users at a completely new level. Not only does it provide the very best in entertainment, and deliver rapid responses to commands and transactions, it also creates a deep personal one-to-one experience between the user and the world first AI genie in the app, Siskabot.

Putting Consumers in the Driving Seat

Siskabot appears in person when the app is downloaded, and she has regular conversations with each user to: 1. Find out the exact type of entertainment content they love, be it live sports, different film genres, series, FashionTV, educational broadcasts, documentaries, and so on. (These conversations are ongoing, as with up to 300 exciting channels, there is plenty to talk about, and we often change our preferences, so that is important too). And 2. To discover what kind of branded products and services they would chose, and the ads they would like to see which could help them save money on anything from cars to household goods to holidays to hot fashion items and beauty products. The conversations are geared to enable OONA to get to know more about each and every customer, and what drives their behaviour. The more the user interacts with Siskabot, the better and more fine tuned OONA's analytics become, and the more advantageous it is for all concerned. - This creates a win-win situation for the consumer, OONA, the telcoms, studios and content holders that OONA monitizes for.

Incentives For Connecting With Your Bot

OONA users are incentivised to regularly link up with Siskabot, and the rewards (in the form of a virtual currency) they earn for these conversations, are on a par with social messaging a friend about specific entertainment they have watched, or would like to see, something they can also do, and get rewarded for, via the OONA app. With the 2017 Telkom Indonesia-OONA partnership, that is set to serve 185 million Indonesians with Free Mobile TV from 2018 onwards, users are given rewards in tcoins which they can quickly build up in a virtual tcoin wallet, and then redeem them for all sorts of things from telcom products to a broad range of branded goods and services, fun days out, and delicious free meals. So with OONA, it really pays to befriend your bot!

In summary, in order to provide super chatbots, the industry and researchers have to integrate multiplex disciplines which incorporate not just deep learning, but also various other technical aspects to achieve a level of technology which merges language, environment and user preferences into one part of flexible intelligent software [1]. Platforms such as OONA Free Mobile TV have come a long way with their research, and deep learning and machine techniques, and it is now just a matter of time before their humanoid bots cross the human threshold. So let's give them a chance!

References

[1]. Mazdak Rezvani (2017). “Why chatbots need a big push from deep learning.” https://venturebeat.com/2017/07/23/why-chatbots-need-a-big-push-from-deep-learning/?platform=hootsuite Accessed 22 Jun. 2018.

[2]. Octane, Al (2017). “What Are The Best Intelligent Chatbots or AI Chatbots Available Online?' Chatbots Magazine.https://chatbotsmagazine.com/which-are-the-best-intelligent-chatbots-or-ai-chatbots-available-online-cc49c0f3569d Accessed 22 Jun. 2018.

[3]. Pardes, Arielle (2018). Wired. “The emotional chatbots are here to probe our feelings.”

https://www.wired.com/story/replika-open-source/ Accessed 22 Jun. 2018.