The Road to (Social) Perdition?

While a lot has been written about Buzz, and I wouldn’t like to regurgitate content, I will list down some key points that I observed while using Buzz and reading the many, many blog posts (like this one) out there. So here goes:

1. First things first: I really think that getting in Buzz as a part of Gmail was a masterstroke. Techcrunch‘s Arrington doesn’t think so, but I feel that for the average Joe, this was the way to go about it. No one wants ‘yet another social network to update’ — other than the early adopters, may be. Arrington argues that Google could’ve rolled out Buzz to a few early adopters, tweaked it and then released it to the wider public. Well, I feel that the best way to get feedback is to release a product to the average user and observe his/her learning curve!
2. While many feel that Google rushed out their product too soon — possibly being in a tearing hurry to get a piece of the Social pie — I feel that this is still in line with their ‘Release-Early-and-Iterate’ philosophy. After all, any feedback is valuable feedback; often, especially the BAD feedback! Yes, there may have been inherent flaws in the product that may indicate insufficient testing. But that again leads to Point #1, may be the testing by techies/geeks/ early-adopters was never going to be sufficient!
3. Buzz product manager Brad Horowitz does not believe that Buzz is a Twitter or a Facebook “killer”.  He says that it’s intended to complement these services. The key to Buzz, he says, is as a conversation enabler. Again, being within Gmail, that is a huge plus.
4. Coming to what I felt while using it, I felt that Buzz lacks one serious feature: the ability to share from within Buzz. Of course, you can share items in Reader (and sharing in Reader has increased since Buzz launched) and they show up in the Buzz timeline, but the entire idea of Buzz was to have ‘One Stream to Rule Them All’! I added TechCrunch on Buzz and reading the posts found myself searching for the familiar ‘Share’ button. Twitter’s “Retweet” functionality (and Facebook’s “Share”) are some of the key drivers to the viral sharing of links on the Web. (Update: Well, I got this a bit wrong, may be. Buzz does allow you to email a post, so technically that’s sharing. But it’s a barrier if I explicitly have to type in names of people to share it with. How about a ‘lists’ feature, which allows you to share with a bunch of folks at a time?)
5. The Buzz timeline takes a little getting used to. Quite a few friends were a little confused about the ordering of posts in Buzz. Another annoying feature is that you get a mail in your Inbox if someone comments after you on a post. Well, if you’re like me, you don’t like your Inbox being cluttered with … well, idle chatter!

But above all this I feel that most commentators are getting one key thing wrong — and that is referring to Google Buzz as a ‘product’. Google would’ve been crazy to launch a major product with so many kinks in it. Did anyone seriously think that Google was in a tearing hurry to launch a major “Social” product without enough testing? I think Buzz is much more about the journey than the destination. May be Buzz is just part of the “big picture” for Google. May be Buzz will change the way email is used. May be Buzz is the bridge between Gmail and Google Wave. Whatever it is, we’ll know in the near future. But, Buzz in its current form is definitely not the final ‘product’. Look at it this way: no one would want to disrupt a standard feature like email without first testing waters. At the same time, email needs to evolve. So what do you do? If you’re Google and if Gmail has something like 176 million page views, then you launch a service into Gmail, and over a period of time let it evolve into something that makes conversation more natural. The early adopters may be the first to switch over from Gmail to Buzz, and may be, just may be, the masses will follow! And email as we know now, will be reborn — without the labor pains!

All in all, I don’t think that Buzz is that bad a product — again, if one were to consider it a product. What do you think? It is the road to (Social) Perdition? Or a revolutionary new paradigm?

Mom, What is Two Times Two?

When I was a kid in the forth grade, and about to start computer lessons at school, I remember the boyish enthusiasm with which I asked my parents, “So does the computer answer your questions? If I type in ‘What is two times two?’ do I get an answer that says ‘Four’?” My parents gently told me that one needs to “program” a computer (I had no idea what “programming a computer” meant, I thought that was what they were going to teach us at the lessons), to get it to perform tasks — or generate answers to questions.  Today, after many years of programming the computer, that vision of a forth-grade kid is already pretty much of a reality. And what has made this a reality? Yes, Search — with it’s various flavors: real-time search, decision engine, social search. ”Search” has become a central part of using the computer, and you can type in the question, ‘What is two times two’ and Voila! You have the answer right there for you! I can only imagine the enthusiasm of a forth-grader at typing in questions and getting an answer without having to “program” the computer.

While “Search” existed for many years before “searching” for something became synonymous with “Googling” for something, it was Larry & Sergey’s seminal “Page rank” algorithm introduced their paper titled “Anatomy of a Large Scale Hypertextual Web Search Engine” that was the game changer. And today, 12 years after that comes another work[PDF] from Aardvark titled, “Anatomy of a Large Scale Social Search Engine”, that has can be as disruptive to “Social Search” as the Google paper was to “Search”. I knew about Aardvark from Techcrunch, but had never taken it for a spin. But then I came across their blog post about their “Social Search” algorithm and it piqued my interest. Then a friend who was using Aardvark said that it’s really cool and worth trying out. So I signed up for Aardvark and decided to try out my first query about … what else but food! “What are places that serve good Italian cuisine in Bengalooru?” I asked.  And yes, sure enough, I was getting replies with recommendations ranging from Via Milano to Fiorano

If the last decade was about “Search” the present one is going to be about how to make search more relevant to the average user, or the how to make it contextual. How do I ask for a question and get answers that are relevant to me? That know what I’m asking for, and the context of my query? Semantic search is going to be a big part of this. As I’ve written, real-time search is going to be another challenge in getting search right. But I feel the real game-changer — and the biggest challenge — will be “Social Search”. How do I get recommendations (like the one I tried out) and answers to ‘open-ended questions’? How do I leverage my network, without explicitly ringing up everyone (other than my close friends) and asking them, ‘Hey, what would be your suggestions for a gift for my Mom who’s turning 60 next month?’ The challenge, to use the HP tagline, is to make Search personal. And computing is getting increasingly synonymous with searching — whether it’s searching for friends on a social network, or searching for answers, or searching for suggestions for weekend getaways near your city, or … well, so much more! Google realizes this, and may be that led to their acquisition of Aardvark. (You can read more about it here, and here.)

The challenge with this, of course, is how soon (it’s a question of WHEN and not IF, IMHO) this gets into the mainstream. Aardvark reportedly had 90K+ users as of October 2009. That number is still a drop in the ocean compared to the number that uses “Search” (Google, Bing, Yahoo, take your pic) regularly. The challenge is attaining that ‘critical mass’ that ‘crowd-sourced’ services like Aardvark would need to attain to become really … well, successful for want for a better word. The challenge for any new technology, as I’ve said in the past, is when your Dad will start using it!

What do you think about “Social Search”? If you haven’t yet tried out Aardvark, do it for a spin and put down your thoughts in the comments!

What’s the right approach to new products? Pick three key attributes or features, get those things very, very right, and then forget about everything else. Those three attributes define the fundamental essence and value of the product — the rest is noise. For example, the original iPod was: 1) small enough to fit in your pocket, 2) had enough storage to hold many hours of music and 3) easy to sync with your Mac (most hardware companies can’t make software, so I bet the others got this wrong). That’s it — no wireless, no ability to edit playlists on the device, no support for Ogg — nothing but the essentials, well executed.

Google Buzz launched less than 24 hours earlier as I sit tapping out this post, and I couldn’t help wonder if it’s got the 3 key features that Buccheit talks about. One of the questions that they ask you in an interview goes like this, “Explain Object Oriented Design to your Grandma.” Well, the point of asking that question is to see if you can put across something in your specific area of expertise to someone that’s totally ignorant of that field. However, in the context of Google Buzz, it would be an equally pertinent question, “Explain Google Buzz to your Grandma!”

The Google BUZZword

It’s a masterstroke to integrate Buzz with Gmail – which already has 176 million visitors a month. That’s a LOT of users that get to check out Google Buzz. But the point is, a LOT of these users will be your average Joe-type of users. Not tech enthusiasts, not web developers, not programmers. The key to whether Buzz is a success or not lies in if these Joes and Janes can find the product intuitive enought to use. If they can’t figure out how to use it – or what it’s for – in the first … may be 10 minutes that they try it out, it’s unlikely they’ll bother to come back and use it again. And that is what Twitter is so good at – simplicity is it’s biggest USP. It’s simple – and that is the beauty of it. Google Buzz might have a lot of potential, but if you can’t explain it to your Dad who uses email just to keep in touch with his friends in other parts of the country/world, then it’s not going to help Google with their foray into Social.

What do you think? Have you tried out Google Buzz. Feel free to buzz in your comments

The Report Card

“What’s your super-resolution for the New Year? Oops, what’s your super New Year Resolution?!” was the joke among the team at work. Well, for the less geeky, super-resolution refers to the branch of signal processing that aims to increase the (generally spatial) resolution of signals. And that was one of the areas that we were working on in 2009. And hence the joke … Aah, yes, now I agree – I’m a victim of the ‘shrinking sense of humor’ syndrome. But, I digress …

Well, with the New Calendar Year 1/12th already over, I thought I’ll finally put paper to pen (OK, the digital versions of them) on some thoughts on the last year, and of course, my “super resolutions” for this Year! So here goes: First up of course, something on the past year …

1. I started this blog in August of 2009. The target was to post once a week, and by the end of 2009 I had 16 posts in 21 weeks. So I felt that was pretty OK. Because I’d made a promise to myself, that while I’ll try and post weekly, I’ll not post something just for the heck of it. Nor will I regurgitate content from other sites and post my thoughts on that content.
2. By end of 2009 I had crossed 1000 hits. I hadn’t put down any target for this, so I cannot really decide how well (or otherwise) it was. However, I was disappointed by the comments. 45 comments on 16 posts, or about 3 comments a post. However, when you take the hits into account, that’s pretty dismal. I realize that while eyeballs is OK, but engagement is much, much more important.
3. On a side note to #2, a friend of mine who is a regular reader (or so I kid myself ) suggested that I purchase a domain name and move my blog to the custom domain. I answered that yes, while that is certainly something I’ve thought of, I do not want to do it unless I have engagement on the blog/posts.
4. The content itself: There were times when I found hard-pressed to come up with a topic to post on. As I said, I’d promised myself that I’d only post when I had something (that in my humble belief) was worthwhile posting.

Super Resolution!

Now on to my ‘Resolutions’ for this Calendar Year. Obviously, these resolutions are professional in nature. While the first few are related to this blog, the rest are some targets that I have set myself. So onto 2010 then: First the blog-related stuff.

1. Increase User Engagement: I’ll measure that by the number of comments. While I do not have any hard targets, my aim would be to get at least 5X comments as I did last year.
2. Again, renewing my commitment to post weekly, I’ll try and have something to share with you all every week. I realize I’m woefully behind schedule, posting nothing in January. So well, that’s something to cover up on …
3. I’ve also thought about posting more often, may be twice a week on … well, rather mundane topics … I read this excellent interview of Mark Pincus of Zynga, and he talks about having 3 targets for a week. I am thinking of taking that up and may be posting on my 3 targets and how I’m measuring up to them. So while they may not really make interesting reading for the world … I intend to use that as a journal to chart out my own progress!
4. Long posts are a killer. While I personally like long posts, I’ve often noticed how while reading other blogs, may be half way through a particularly long post I start skimming over the article looking for that interesting bit. So those of you who are bored by the long posts aren’t going to have a total reprieve, I promise to shake things up a little bit.

And now something more general:

1. Learning Android. In my opinion (and well, you don’t have to be a genius to have this opinion), the Android is THE platform of the future. I’ve been picking up Android basics since about December last year. While I won’t say the progress is as fast as may be I’d have liked, it’s been there. I’ve set up a target to be fairly familiar with Android development by April this year. And oh, did I mention building an Android application by the halfway mark for 2010?
2. Focus, Focus, Focus. Yes, one reason why I feel the pace with the Android learning curve has been slow is because I haven’t really given it my 100 per cent. I can of course say that I didn’t know too much Java, or that I was learning Android as a 20 per cent project, but well … let’s stop the buck-passing, shall we?
3. Getting out of the comfort zone. This is more vague, but I’ve decided that I’m going to get myself out of my comfort zone and pick up on new things. Explore new avenues. Look out for interesting applications of technology in everyday life. Think out-of-the-box. OK, I’ll stop with the buzzwords …

So, that’s it … Pretty much these are my Super New Year Resolutions! What are yours? Let everyone know in the Comments!

Robot Vision

So the sub-plot that’s being played out is that the leading lady Nisha (played by Madhuri Dikshit) is getting married (pretty much against her wishes, but to keep the family happy blah-blah) to Rajesh (who’s been widowed and has a new born kid to take care of). Now Nisha loves Prem (Salman Khan, who manages to keep his shirt mostly on) who also loves her. So Nisha’s decked up for the pre-marriage ceremony but overwhelmed by sadness. Nisha gets Tuffy, the lovable family dog, to deliver a note and a necklace that was gifted to her and Prem by her sister as the lasting token of her love.

Now, if you’re still with me, starts the fun part (and the subject matter of this post). The dog bobs along and reaches the hall where the ceremony’s about to begin. He sees Prem happily playing with Rajesh’s kid. And Rajesh is getting ready for the ceremony to be married to Nisha. And the camera shows Tuffy look alternately at (what the viewers know, of course) are Rajesh and Prem. But finally he jumps up and gives the envelope and the necklace to Rajesh. And then of course, the film ends happily with Prem and Nisha united in matrimony.

What interested me in this sequence was the visual system and visual understanding and the role that context plays in it. Now of course, it was the dog’s visual system (as opposed to the human visual system) that was in play in the movie, but I feel that it has a parallel in primitive or early attempts at machine vision. The canine visual system is a ‘primitive’ visual system that perceives shapes but is not equipped with the sophisticated face recognition that we humans are endowed with. However, the context plays an important role in identifying people or understanding situations.

The Eye

I have a question for you all. But before that let’s set the stage for our little act — let us replace the dog with a humanoid with a primitive visual system but let us say our humanoid can learn from situations and extract meaningful, if simple, concepts from them. So our humanoid sees a human playing with a baby and another human decked up to attend a ceremony. It also sees a female decked up for the same ceremony, who hands it a message to be delivered to a specific human. Now let’s assume that the “World” has just five primary characters: the four humans — Rajesh, Prem and Nisha, Rajesh’s newborn son — and the humanoid. Now the humanoid has observed interactions between the human characters amongst themselves, and learned simple concepts of relations. So the humanoid has learned the relations or associations that link the baby and Rajesh, Prem and Nisha. It has also learned the concept of a bond between Prem and Nisha. However, it doesn’t have a sophisticated face recognition engine and also doesn’t have a very … evolved brain that we humans have. Now, if given this context, it has the scene that I just described played out before it, who would it pick as Prem? That is, would it really pick (who we know is) Prem or Rajesh?

What I was trying to understand is the role of these learned associations in visual understanding. Can we “recognize” people or situations from learned simple concepts and associations aiding a (relatively) simple visual processing system? As a computer vision technologist, I’ve studied the role of context in visual scene understanding. However, here’s a real-life (well, reel-life if you please) example that kind of puts to test our understanding of the human visual system!

What do you think about this?

Eye to Eye

What do you get when you put together a geek who fiddles around with computer vision in his day job, a movie with some cool 3D  SFX, a 15-odd hour bus travel and the in-travel movie? Well, if you’re me, you get this post! The first part of this post is going to be about the experience that was Avatar. Yes, Avatar is an experience more than a movie, and if you’ve gotta watch it, watch it in 3D or not at all. Mike Arrington called it the iPhone of movies. I’m not sure I quite agree, but the SFX are pretty compelling. But as with any technology, I think as technologists we’d do well to cast aside the technologist in us for a moment and think from the average Joe’s point of view. One of my friends that I was talking to after the movie said that the 3D experience helps us ‘focus’ more on the movie. I’m not quite sure what he meant, but he probably meant that the depth perception aids the overall experience. That set off two thoughts in my mind. First, was the question that I asked him: “You probably feel it because you’ve worked in the multimedia technology field for almost a decade now. But would your wife (who’s not a techie) feel the same?” I believe that the answer to that question would be the key to the mainstream adoption of 3D displays. I guess we’ve got so used to experiencing movies in 2D and the rest of the (real) world in 3D that we’ve probably conditioned ourselves to the difference in the two. That is, the consumer doesn’t quite realize that the 2D movie experience is different (and lacking) from the way s/he perceives the real world around! And I believe that that’s the key: giving the user something that she didn’t even know she wanted till she has it. In that sense, may be, the 3D experience can be the iPhone of the entertainment industry. Avatar is compelling, but the experience can be a lot better.

Avatar - 3D

How could the experience be better? The key factor is how the content is consumed. The 3D glasses (anaglyphic or polarized) are presently quite bulky. And if you’re gonna watch a full 180-minute movie, or a 90-minute football game in 3D, then that’s definitely going to be a factor. I also felt that the experience can be more immersive, in that the user should feel that she’s right in the thick of action. Somehow I couldn’t get that feeling while watching Avatar. Is it possible? Well, I don’t claim to be the expert here, but how about giving the user the choice of a viewpoint from which to watch the whole thing? It would be great to watch it from the viewpoint of Jake as he dives on Toruk, the giant flying bird. Now that would be something, isn’t it? If you can watch multiple camera angles during a football game, then I guess this should also be possible. I know for sure that there are patents in this space: rendering novel views from a few captured views. So may be companies are already working on it! The iPhone revolutionized the mobile phone space because in its design and UX it threw out of the window all past experiences on the mobile phone. I think that the 3D experience can do the same thing to the entertainment industry. But is Avatar the experience that we want? Well, not quite, but it’s getting there, for sure.

Update: Here’s the next post in this two-post series

JaMocha Tech

JaMocha Tech is a venture started by Rajesh Sharma, and looks to deliver employee productivity analytics to enterprise customers, through their ProHance solution. ProHance is based on the NetValence Framework, and you can find out more about it here.  Their main target is the BPO segment, where they say that they provide value through delivering solutions at a fraction of the cost of their competitors. They also focus on an intuitive report generation/graphs showing the captured analytics to the user as their USP. Rajesh and his co-founder Kishore come with some very solid experience and insights into the way the segment they’re targeting works; so that should be a real plus. Understanding what your customers wants, and sort of having been-there-done-that is a huge plus. Truly automatic monitoring may be difficult in a general scenario, but the IT/ITeS or BPO scenario is different, and that’s what JaMocha Tech is claiming to handle really well!

Rent Toys - How it Works

The next startup to pitch was a really innovative idea. You’ve heard about renting out books, CDs, sports equipment … but kids’ toys?! Well, that’s what Rentoys aims to do! It provides rented toys to kids (well, it deals with the kids’ parents, of course) for a monthly subscription, with plans beginning from 199 bucks. They do a delivery of the toys right at your doorstep! I would think that that’s kind of difficult to do, but they’re in talks with a couple of courier companies to handle this. Right now, they’re doing it themselves! I think that the segment they’re tapping into has a lot of potential, and I also think that they have the ‘early-mover’ advantage: it’s pretty much something that no one’s attempted to do so far! In his pitch, the founder (apologies, I don’t remember the name, and the website doesn’t have one either), Manoj Kumar, talked about how kids get bored of a toy after a few weeks or even a few days, and so it’d be a nice idea to rent out a toy, rather than buy it. On a lighter note, I wonder how they’d handle the reverse scenario: the kid loves the toy so much that s/he’s not willing to part with it!

One thing that strikes about the RentToys venture is that it was the only ‘non-tech’ company that pitched at the meet. Yes, Sports keeda can be called non-tech too, but it’s a Web company/venture (involving team of bloggers). And I remember mentioning to my friend who’d accompanied me that sometimes we techies get a bit too technical! I think there’re real world problems waiting to be solved, and may be sometimes we should just keep our tech glasses aside and look at the little things out there!

I would like to continue on this series with a few thoughts of my own from the interactions at Start Up Saturday meet today. So watch out this space for more! And as always, feel free to leave your thoughts on the start up scene in Bangalore/India in the comments!

Just Do It!

Just do it. Go on, put that idea that’s been lurking in your mind to action. Be a scrapper, and don’t worry about the naysayers. At the same time, do take feedback from your mentors, friends, beta testers, and be flexible enough to change course as the situation demands. Those will probably be the takeaways from today’s Start Up Saturday meet up in IIM-Bangalore. I came to know about SS (Start up Saturdays) during a pretty freewheeling conversation about ideas, innovation, entrepreneurs, and start ups. And I was pretty excited to be attending my frist SS meet up and it was well worth driving across the city to be a part of the meet up. So what I’m going to do is write about some of the lightning pitches that were presented and a little bit about the panel discussion that followed.

InterviewStreet

First up was Vivek Ravishankar presenting his Job Interview service InterviewStreet. InterviewStreet (@interviewstreet) targets mainly the fresh-graduate segment or students who’re applying for job interviews. It tries to help them prep up for the placement process by setting up mock interviews with people in the industry. Vivek said his USP was that they can get feedback and an accurate idea of where the bar (for a particular employer) is set, and where they stand and this will help them improve. InterviewStreet makes money by charging the students who register for the service – 350 bucks for a single interview session. I have two major gripes with this:

• First, I’m not sure students would be willing to cough up that amount for an interview session with different industry ‘experts’. Students would typically be having contacts (alumni, friends) in the industry who, if not conduct a mock interview, would be willing to share notes on how to prep up for the interview process, what to expect, etc.
• The second gripe is that I’m not sure that it doesn’t constitute a ‘conflict of interest’ for the interviewers who register with the site. The site would be paying them for the interviews and that’s like making money off your employer. Well, the employer is not going to be happy to know about that.

That aside, I think it’s a pretty novel concept and may be worth watching. I would like to see the concept evolve from just setting up interviews to may be a more customized offering for a student – feedback from the interviewer and tips to improve specific areas: not only technical areas but also soft-skills. The market that they’re targeting is pretty huge: about half a million students graduate out of college every year. Even if they convert a fraction of this base for starters they would be do pretty well!

SportsKeeda Home Page

Next up were Porush Jain and Robin Chabbra demoing their start up SportsKeeda - a group blog on sports. Porush and Robin are final year students at SIBM. Their aim is to be the go-to site for sports enthusiasts; where the contents comes from the enthusiasts themselves.  A look at their website shows a lot of content contributed by different bloggers. I am not sure how they plan to monetize the blog – obviously advertising could be one way, but I guess there could be other (and more exciting) ways as well: how about selling event tickets? (I don’t know if that’s possible in India, but just thinking out aloud.) I guess sustainability is going to be the key here – I know from personal experience that writing regularly is easier said than done! I asked them during the Q&A round if they’ve explored the possibility of going local – or even hyper-local – and writing about university/club/state level sports events. One key thing that helps a sportsperson is getting recognition at the junior/state levels. May be they could have a “Featured Sportsperson” section every month. That would involve a lot of travel and meeting up and talking to people, but I guess that should be an interesting point.

To be continued …

You can stay up-to-date with SS (and other HeadStart news) by following @HeadStarters on twitter!

Update: Here’s the next post in the series.

Discovery vs. Search

The title of this post is an amazing collection of short stories by Richard Feynman (of The Feynman Lectures in Physics fame). If you haven’t read it, I would seriously recommend you do. The book is typical Feynman, where we get a glimpse into his innate curiosity and way of searching for meaning in simple things — from the cutlery on your table at the restaurant to the line of ants ferrying small sugar particles across the room. That pleasure — of finding things out — is also applicable to the Internet. Given the humongous data that the Internet puts at the disposal of a mouse click, the question is, ‘How do you discover interesting things from this mountain of data?’ It’s the proverbial ‘finding a needle in a haystack problem’.

Think about it for a second. Everyday, every minute that we’re online we’re having information literally thrown at us by old media and new — news stories, blog posts, tweets, op-eds, videos uploaded on video sharing sites like YouTube, Hulu, live-streaming sites like Ustream. How do you surface information that’s interesting to the user — that she would want to consume, but yet doesn’t know is out there. Let’s say you’re an avid follower of mobile news. So how do you get a digest that crawls the huge WWW and returns interesting stories (blog-posts, product reviews, new launches) for you? The key, as in Feynman’s book is the Pleasure of Finding Things Out.

Search rules the Web today, but soon it’ll be about surfacing content, or discovering it, rather than searching for it. You know what you’re searching for, and just want to find out where it exists on the Web. What do you do if you do not specifically know what you’re looking for, but want to find things that you’re interested in. You know, the latest apparel trends for a budding designer, the latest stuff on product launches/technology, or just videos that you’d be interested in. (Check out this post about VidQue, that aims to surface popular videos based on what you/social circle has watched.) What we need is a Discovery Engine rather than a Search Engine! Thus far it’s been about data. The data was the center of everything. The next decade will not be about data — but about how to find the data that you’re looking for, before you even know that you’re looking for it. (Like Sergey Brin says, the ideal search engine begins with you thinking about something, and it knows it.) It’s also going to be about analytics. Data by itself is just that, bits and bytes of information. What is going to be important is having analytics for that data. Of course, analytics has been around a while, but the key is making it pervasive. What can you learn from data-access patterns, how can you put (what you learned) to use, and who can benefit from it. It’s going to be about finding data for the user, and about tracking data-access for the provider. Tracking data in turn enables better content discovery for the user, and enables her to sift through the mountains and get to what she would want!

Of course, the problem is a difficult one, but that’s what makes it exciting! It is similar to what I talked about in the previous post on content-based retrieval. The common thread is about bridging the semantic gap. Can we bridge the semantic gap between the user’s understanding of something (or what she associates with information/data) and the meta-data that’s attached to the piece of information? Bridging the semantic gap will be a key thing to solve the discovery problem — you need to discover topics/subjects that the user is interested in (from access patterns) and present the information to her.

I’ll keep this post short (and a little rough around the edges) deliberately. May be I’ll update this thought process as it matures a little bit. But I also wanted to keep it that way (a little rough) to spur discussion. Feel free to put down your thoughts in the comments!

Finding a needle in a haystack?

Welcome to the world of content-based image retrieval or CBIR. CBIR is the science (well, some would say it’s an art) of querying a collection of images and videos using features (or properties) from the query that you’re presenting the system. It is also sometimes called visual search — though visual search also refers to presenting the results for what you’re searching more visually - rather than the text/link interface that we’re more familiar with. Image retrieval remains a challenging field, primarily due to what researchers call the semantic gap - the gap between a human’s understanding of image content and the properties that a machine can compute from the same image. My graduate thesis in IIT-Bombay was in this domain of CBIR – where I worked on the problem of scene categorization. This week I am giving a short talk at work on CBIR, and I’d like to share the slide-deck of my talk with you all.

I would be glad to answer any questions that you may have, so feel free to put them down in the comments. Or drop me a line at prabhudesai[dot]amit[at]gmail[dot]com.

Update: Added the slide-deck of my graduate thesis presentation – this is also a part of my talk.