Last week saw IBM testing its latest supercomputer – Watson – in the US game-show ‘Jeopardy’. Pitted against the show’s all-time reigning champion and the record holder of the show’s longest winning streak, Watson crushed its opponents in the four-day contest to win the $1m prize.
The face of Watson, designed by Joshua Davis
To briefly summarise, in IBM’s words, “Watson simultaneously runs natural language processing, information retrieval, knowledge representation and reasoning algorithms to fathom the intent of questions and yield what it thinks is the best answer”. In plain English Watson understands the meaning behind questions, draws from a vast pool of information and delivers its answer in a meaningful way – returning knowledge not just data. While ‘Jeopardy’ may be the perfect way to test its abilities (I believe previous IBM supercomputers played chess!), thinking about the implications for the future of online search engines is mind blowing (if you’re into that kind of thing…).
Well, even if you’re not, we all use Google as a first port of call when we want to look something up. We type our keywords in and get a list of pages which feature those keywords most frequently. This may be over-simplifying it, but there’s very limited understanding of meaning in Google’s search returns – imagine your searches being understood in the natural language context of your thoughts…
IBM is working to develop this technology into a commercial product over the next couple of years – most likely focusing on medical applications but it’s easy to see how this kind of technology will eventually filter down to mass-market consumer applications.
Also recently featured in the news is Wolfram|Alpha, an online answer engine that performs calculations in real-time to serve a host of detailed, usable information. Although currently focused on performing scientific and technical calculations such as “boiling point of water at 6 atm” or “lim(x->0) x/sin x” there are many practical things is can do to help in every day life such as calculating mortgage repayments and weather predictions. Wolfram|Alpha will return written and graphical information and data tables by understanding the nature and context of the query, cross referencing different sets of data and delivering calculated responses. Microsoft’s search engine bing is already integrating Wolfram|Alpha into its search returns giving users immediate calculated responses amongst its key word-based returns.
As we see more sophistication and complexity in the channels we use for information, so too will we see much richer, more intuitive user-experience where the human-computer interaction will mimic the way we think, talk and process information – with the storage and processing of a world of information.
