Wednesday, October 11, 2006
Interview with W. Lewis Johnson, Founder of Alelo
We recently ran across Alelo (www.alelo.com), a startup and University of Southern California spinout developing interactive games used by the military. The company's very engaging, interactive 3D role playing games teach languages like Arabic and Pashto to troops being deployed to the Middle East. Using speech recognition and other technology, the titles teach foreign languages to players as they go through the game in simulated environments like Iraq. We spoke with Dr. W. Lewis Johnson, CEO of Alelo, about the firm's technology and plans.
Ben Kuo: Tell us a little bit about Alelo, and what the company and product does?
W. Lewis Johnson: Our overall company is called Alelo, and we also have a government subsidiary called Tactical Language Training LLC which develops projects for the government. We develop interactive products for teaching foreign languages and cultures using video game and artificial intelligence technology. We're a spinoff of the University of Southern California, where I am a research professor in Computer Science.
Ben Kuo: Tell us a little bit about where the company came from?
W. Lewis Johnson: The company originated out of research project out of USC called the Tactical Language Project, which was funded by DARPA. The research project was a a response to a solicitation by DARPA to develop an innovative approach to teaching basic communications skills to wide range of people in the military. The purpose was to break out of the mode of the past, where there were small numbers of highly trained linguists, and to distribute language skills very widely. We started work on the project in early 2003, with a prototype of tactical Eastern Arabic. We gave it to the U.S. military for evaluation, and quickly discovered in the Fall of 2004 that Iraqi Arabic was the major need. We started working on an Iraqi Arabic prototype, which we delivered to the Marine Corp. for testing in the middle of 2005. At about the same time, we set up the spinoff company. The reason for that was there was clearly a big demand for this, and the university was not equipped to provide customer support, training, and further product development. This enabled us to progress beyond the research stage, and I also saw lots of potential with the technology. My goal here was to create an approach to learning foreign languages that is substantially better than what is out there, and to help people across the board to learn languages and cultures very rapidly. We were very encouraged by the prototype, and set up the company in order to pursue those market opportunities.
Ben Kuo: For those who haven't seen the game, can you describe how it works?
W. Lewis Johnson: So the way our product works, is that it is built on top of a commercial video game, the Unreal engine from Epic games. Basically, how you deploy it is on a PC, where you plug in a digital headset microphone, and using that and the mouse you can enter the virtual game world we create. You can enter into a conversation with characters in the game, and you can select non-verbal communications in addition to verbal communications. For example, you can use the mouse to select hand gestures as you speak, and you can see how characters respond to that. We couple that with a set of interactive exercises that are built into the game, to help you acquire skills you have to play the game. Its very task oriented. Basically, everything is focused on learning what you need to know to carry out a particular task or mission. Since last year, our company has been developing the engine further, and obtained a source license for the Unreal engine and integrated our technology with that, so that it's a single program that you download and install.
Ben Kuo: Can you tell us where your software is being used?
W. Lewis Johnson: We started working with a number of early adopter sites in the military, one of the first was the Marine Corp. Expeditionary Warfare School in Quantico, Virginia; the First Marine Division in Camp Pendleton, and also with Fort Riley, in Kansas. It's really only in the past month that we have had a version of the software which is well developed and being distributed widely. We have now distributed in excess of 1500 copies of Tactical Iraqi, and the military has been making their own copies and distributing them, as well. We don't know exactly how many people have used the software, but I know that several sites have hundreds of people who have already gone through training with it, and the numbers keep adding up. We've reached an inflection point where enough people in the military are taking notice of it, and are seeing what people are doing with it, that units want copies for themselves. We now have it available for download to everyone in the Marine Corp. There's strong interest in the Army and also Special Forces, who have been great supporters. We also continue to get support from DARPA--not financially, as we've gone beyond what DARPA funds--but in terms of help in contacts with people in the military that are potential users. A couple of weeks ago we had a booth at Wired Magazine's NextFest at Javitz Center in New York, and we had hundreds and hundreds of people coming through from the general public trying our systems out. Everyone got it, and the value of it. Anybody over the age of 10 understands what it is and what to do with it immediately. It generated lots of excitement.
Ben Kuo: So what are your plans from here, it looks like you are getting good military adoption, is there any plans to go beyond that?
W. Lewis Johnson: We plan to continue to develop products for the military side, and at the same time to also develop versions for the general public. On the military side, we have a beta version of our Tactical Pashto trainer, for Afghanistan; and we also have a French version for sub-Saharan Africa. We also have requests for other languages. At the same time we are spinning up titles, we are going to start distributing to the public--one for Gulf Arabic, for businessmen who are going overseas, and also one for French, focusing on Metropolitan France. We expect to be developing additional products of those sort. We're also exploring possibilities of licensing our technology to other people interested in developing similar products. We already have discussions under way. We also see the potential of using the technology for anything that uses communication skills and task skills. For example, you could apply this to a wide range of job training applications, an example would be training customer service representatives, working a front desk at a hotel, or for healthcare professionals who need to interview patients possibly in a foreign language. We think that the potential markets here we can tap are quite large, including language training, cultural awareness training--both in business and government--and other simulation task training applications.
Ben Kuo: Where's the idea come from for using games to teach a language?
W. Lewis Johnson: I have been doing research for several years in what we call pedagogical agents - animated characters that can act as guides and tutors. These are things like the Microsoft Clippie but done right--which can understand what a person is going to do, and be in a better position to help. We applied this in a number of different areas, including medical training and engineering skills training, but it's been in the back of my mind that language training would be a natural sweet spot for this type of technology. That was one thing, another thing was that game technology--specifically, 3D game technology -has progressed to the point that we could take advantage of building these products. I started about 10 years ago working on using virtual reality for training using head mounted displays -- and it was kind of a cumbersome way of doing things. Since then, the game engines keep getting better and more powerful. I came to recognize that mod-able game engines--which users can modify the game world--are quite effective tools building on top of these kinds of platforms. Some of the students in our lab at USC had taken the lead in integrating AI technology with these game engines. We thought let's take this and build on this and build a robust language learning environment. We didn't know if we could pull it off, given the nature of speech recognition and AI, but it worked pretty well. The Turing Test, which is creating an animated character you can engage in real conversation is very hard. But, in context of a concrete foreign language, where there are a limited number of things you can say, it's well within the technology.
Ben Kuo: Can you talk about the user experience with the language recognition? I was impressed by the fact that the software recognizes what you say in Arabic.
W. Lewis Johnson: They speak into the microphone and the interaction is different depending on the part of the program they are in. If they are in the mission game and talking with animated characters, and they say something in Arabic or French, the character will respond in that language. Generally what those characters in the game are doing are to understand the gist of what you are trying to say, and respond accordingly. In the interactive exercises we have a range of different exercises that give you feedback on your pronunciation, use of words, and grammatical forms, to build skills that you are going to apply to interactive game portion. When you are in the game world, you might be in a cafe, on a street corner in Iraq, and you can walk up to a character and speak to them. The speech recognizer is looking for what you might say--you might introduce yourself, you might ask some questions, or greet the people there--and recognizes what you are saying, and the AI is trying to figure out how to respond to that in context of the dialog. There's a dialog model which indicates what the appropriate responses to what a user might say in a different context, which leads the AI to choose the proper response to drive the animation and non-player character, and the choice of what a character says.
Ben Kuo: How has the spinout experience been for you?
W. Lewis Johnson: The process has gone exceedingly well. We have gotten good support form the military, from DARPA , as well as from users. We have also had good success in attracting government funding of both research and development as well as the technology and content development. We're now working at a multi million a year rate. The university is now a shareholder in the company, and although it took a while to negotiate the final agreement with them we now have that in place, the university has been quite receptive to have this kind of thing go on. In that respect it's a big change from what I observed in the past. When I first came to USC, the institution was rather suspicious of faculty members starting spinoff companies. At this point they have a much more positive attitude on it. There are various challenges to overcome as a startup, but so far we've been managing growth pretty well. In the next year we'll be ready to take on investment to energize the commercial phase of the development even more. I'm very hopeful for the future.
Ben Kuo: Thanks, and good luck!