Advances in artificial intelligence (AI) and virtual reality (VR) are stealthily shaping the world around us.
From facial recognition systems that understand human faces in photos to VR headsets that put you in a simulated Tour de France, AI and VR are integrating into our lives and becoming increasingly common technologies utilized by some of the country’s most forward-thinking companies. Boston is home to a number of startups doing big things in the AI and machine learning field and we spoke with a few of their leaders to hear about what they’re doing, and where they think the field is heading.
Response from Rana el Kaliouby, CEO of Affectiva
Affectiva, an MIT Media Lab spin-off, is a pioneer of Emotion AI, one of the next frontiers of artificial intelligence. The company is bringing emotional intelligence to the digital world through its emotion recognition technology that senses and analyzes facial expressions and emotion. Today, Affectiva is used by more than 1,400 brands to gather insight and analytics in consumer emotional engagement.
What are three of your predictions for AI and/or VR in 2017? What about three years from now?
We are quickly seeing a de-facto engagement mode of how we interact with devices around us. An example of this would be in Siri, or in the proliferation of similar chatbots. Emotion AI plays a very important role in that kind of interface because it needs to have some basic emotional or social skills in the way it engages with the user. We think this will start to really manifest this year.
We expect the conversation interfaces that are able to develop this social rapport element right are going to be more engaging and sticky, and result in less frustration for their users and see a proliferation of the technology.
We also see these conversational interfaces showing up in more of our everyday lives, and think that Emotion AI will be on a quick path to ubiquity this year and these will range from ordering a pizza to enhancing your driving experience in your car.
What about 3 years from now?
We a unique perspective on this, and see the AI space moving very quickly. AI will become more conversational and relational in its interfaces and interactions with us.
All of our devices have GPS, or have some sort of location-enabled capability. Maybe three years from now our devices will have an emotion chip that will read facial expressions, analyze tone of your voice, have built in emotion awareness to adapt our experiences to our emotions in real time.
What are the biggest challenges the AI/VR field(s) face right now? How are you overcoming them?
The AI field has made so much progress because of two main factors: algorithms/computing, and data. At Affectiva, the next leap in AI is that by investing both of these components (we just hit 5 million facial videos!) we can better leverage the data — because the biggest challenge for us is that we are amassing data at a rate much faster than a rate than we can actually make use of. Incorporate process into building new emotion classifiers. All of machine learning we do so far — supervised machine learning approach (adult informing learning process) labellers, we give them the data, then algorithm that data. We are currently bottlenecked by how fast our human labellers can label the amount of data we have to feed our algorithms.
We have a great opportunity is to break new ground in unsupervised machine learning. Currently, all AI companies use semi-supervised machine learning, where the process is that we teach the algorithms a little, the machine learns, we check in, etc. so the algorithm is continuously improving. The next big leap for us in AI is when we have unsupervised data. With no human in the loop, we’ll lob data at the algorithm directly.
What is the most rewarding part of working in this field?
Personally, for Affectiva, it was starting a new industry — Emotion AI — that we are continuously working to tell the story of and define. We came out of academia, and are projected to grow (we are in a high-growth market with lots of commercial opportunities, and excited to be a part of that).
Another piece of how we interpret rewarding work is what we are able to do with our partners, such as Hirevue (emotion tech to help employers understand who to hire) and Cloverleaf (better leveraging consumer emotions in retail). And as a general rule, this type of emotion tech has a huge opportunity to make a dent in the mental health space, such as our partners at Brainpower who are leveraging our tech to help those with autism better understand emotions.
What advantages do AI/VR companies have being located in Boston?
Timing is great on this as we actually officially moved to our new office in downtown Boston from Waltham last week! Being in Boston places us at an intellectual center with MIT and Harvard in our backyard: in fact, Marvin Minsky, the father of AI, came from MIT. As a company, we are an MIT spinout, from the Boston MIT Media Lab, so we are close to that community and close to that research. Boston is also a big healthcare and information Hub, and is rich with tech companies and startups — so we knew we would be in good company. It was a strategic move as well as we look to growing the company with new hires — and a Boston location will help us to attract top-level talent.
Responses from Brennan White, CEO of Cortex
Cortex is an intelligent SaaS platform that helps brands make better marketing content decisions. Using marketing content and deployment data, Cortex automates the majority of the decisions behind creating marketing content, promoting marketing content and monitoring competition — helping brands compete on mass marketing channels.
What are three of your predictions for AI and/or VR in 2017?
By the end of 2017, over 50 percent of major brands will have at least one applied AI (like Cortex) improving an aspect of their business.
By the end of 2017 we'll see the debut of the first full game in VR. To date, most games and pieces of entertainment have been bite size and snackable. The first full-featured game (20+ hours of gameplay) will be done this year.
AI and VR will make up over 20 percent of VC dollars invested in technology this year.
What about three years from now?
AI and Applied AI companies will have near-seamless speech interfaces (like the movie Her). Humans can choose to speak to their work software when it's more convenient or comfortable.
True Roomscale VR: The basics exist now, but precise representations of your space in VR that actively update based on changing furniture will be common. Precise human movement sensors (down to the millimeter) will be present in everyday objects like phones and televisions which will enable VR (and basic AR) to be an option in many homes and businesses.
Image recognition algorithms will be comparable to human ability in some cases (and much better than humans in many cases). Persistent, visual safety checks for building safety will begin being deployed on bridges and other structures.
What are the biggest challenges the AI/VR field(s) face right now? How are you overcoming them?
The data for most AI problems is very noisy. Cortex is building models to clean or expand the data set to move beyond these problems.
What is the most rewarding part of working in this field?
Building something that is relied upon by global brands across departments, timezones and languages to improve their thinking. Seeing the positive career impact on those who dare to embrace new technologies like this to improve their departments is a massively gratifying thing.
What advantages do AI/VR companies have being located in Boston?
Boston has the right mixture of AI and ML PhDs and students and the right investor and technology community to be building several exciting AI/VR companies simultaneously. Beyond that, we have the support of the city to build globally-competitive businesses.
Responses from Scott Evernden, CTO of Infinite Stores
Infinite Stores is a platform for making 3D and VR shopping experiences. The system allows an individual or group of artisans selling products online a way to occupy a virtual store within an endless shopping universe that they can fill with items and visually merchandise — just like a real-world store.
What are three of your predictions for AI and/or VR in 2017? What about three years from now?
Regarding VR, I think my main predictions might be the following:
1. Google Daydream devices capable of quality VR become more widely available which will increase VR’s audience and most importantly grow awareness of this new technology. We may see positional tracking for these head mounted mobile devices.
2. Microsoft will enter the marketplace forcefully offering a head mounted display for half the price of the current high-end offerings. This is certain to advance the audience and general awareness too.
3. Really great virtual reality games and entertainment will be produced. But VR will start to see uses in other areas including social, educational, travel, medical and e-commerce.
What are the biggest challenges the AI/VR field(s) face right now? How are you overcoming them?
Engagement. Many VR experiences can be very immersive, fun, and entertaining, but it’s a challenge to have your users returning hoping to find something new. This is especially evident in many VR store examples I look at that simply do not change from one visit to the next.
Content. Like everything else on the web, VR is driven by content. Someone has to create this content and it’s usually not the programmers, so artist and creatives should start considering developing their valuable skills in this area.
What is the most rewarding part of working in this field?
I think anyone making things in AR or VR is being rewarded with valuable skillsets that will serve them well for a good long while in creating a wider range of great new applications and really innovative experiences. I feel lucky to be able to witness and hopefully be a contributing factor in migrating us into the metaverse.
What advantages do AI/VR companies have being located in Boston?
There is a rich talent pool in Boston given the large number of both technical and creative colleges/universities in the area. VR needs artists and creatives more than any previous IT technology. Content is king to the 3rd degree as our man-machine connection evolves from words and pictures, into worlds and fixtures, and avatars and architecture, environments, wind and smells. Someone has to make all of that. The many technical institutions here are right now preparing a new breed of spatially-oriented programmer who’ll conceptualize and realize some amazing uses for this new technology.
Photos via social media
Know an awesome startup in the AI/VR space? Let us know or tweet us @BuiltInBOS
