wtorek, 22 listopada 2011

Future speculations

The future of the idea is discussed annually on  International Workshop on Adaptive Multimedia Retrieval.
This meeting gathers experts and passionates for the subject to speculate on possible inventions and perspectives. Examplary topics of discussion:

  • Visual Search Performance of Interactive 2D and 3D Storyboards
  • Image Classifying
  • Concept of omnipresent multimedia
One of the possible future usage is described in this article:
http://www.ercim.eu/publication/Ercim_News/enw62/jose.html

The possibility to search content in images , video or music isn't developed pretty well yet, but i suspect, if concept will be developed, it will drastically improve searching media abilities.

The influence on every day's life, as with most of the computational subject, will continuously grow.

Adaptive multimedia and it's retrieval examples

In every day's work, or fun, we are using (or playing, what would be explained later) adaptive multimedia systems. The topic is really huge, ranging from very simple and old systems, to a very complex and futuristic ones. I will try to show, maybe to your surprise, few of the examples:


1. 

That's right, the most popular video site nowadays (although some says, that it's getting old technologically).

This is a very standard example, as it offers a multimedia adaptive system to a huge amount of people worldwide. So what are the adaptive YouTube elements?


  • The ability to comment, leave a rating or favorite a video. Rating system give us the power to show other users, that were searching for similar subject, at first video that we think will suit their needs. That's one of the steps (a small one, but always) for creating a semantic network, with a possibility to highlight best results regarding the topic, and bury worst or out-of-topic videos. Comments creates a social network element, when except sound and picture, social exchange of thoughts comes to play. This ability let's us start a discussion, like for example on a forum, but with topic presented in a form of a video. Another social network element is favoriting a video. When video is favorited, if we have a youtube account connected to facebook account, all of our friends will see favorited video. That gives us a possibility to share interesting picture and start a discussion about it.
  • Related videos section. This feature lets us know more about the subject that we were requesting, creating a chain of videos with the same or similar tags (keywords describing the video).
  • The ability to put a note on the video, explaining something that is happening right now on the video and showing exact location of it.


2.
Rock Band

Games from the very beggining offered multimedia experience with picture and sound, and required a gamer to activate his senses and use input device to steer the game. Although arcade machines existed some time ago, where we had a steering wheel for driving game, or a skateboard connected to machine for skateboarding game, most of the games used standard input device : keyboard, mouse, joystick or pad.
These devices didn't gave gamers opportunity to emulate in real life what was happening on the TV/computer screen. This have changed with games like Dance Dance Revolution, Guitar Hero or aforementioned Rock Band. These games, with their special devices gave that opportunity to emulate things done virtually in real life. The adaptive part of it all is, depending on our skills in playing a controller-instrument, playing the music more or less accurate. So, when the player is not doing well, the music played back to him is sloppy. On the other hand, with a good player, he can play a song perfectly, and so does the game responds with clean and undistorted song coming from the speakers. The social aspect is, that when played with friends, everyone is responsible for the result. If one instrument is played badly, the game responds with a sloppy tune. The next generation of those kinds of games involves motion sensors (Kinect, Move, Wii) that doesn't require any additional controllers and let's us fully control the virtual character with the player's body moves. This is a one step closer to fully virtual realities when we can stimulate any sense by multimedia.


3.
Siri is an intelligent software assistant and knowledge navigator functioning as a personal assistant application for iOS. At least that's what Wikipedia says.
The idea of voice-steered systems is old, but, as i discussed few weeks ago with my teacher, it never really gain success. We weren't sure why is that, the thing is, Apple is trying to change it. And the might succeed with the idea. The software stimulates speaking and hearing senses, and that wasn't seen often  in mobile devices to this point. What's the possibilities of Siri?
"Siri claims that the software adapts to the user's individual preferences over time and personalizes results, as well as accomplishing tasks such as making dinner reservations and reserving a cab."

As we can see, the adaptation element is present in the software and it touches many different aspects of humans life. 



With these examples i was trying to show a wide range of usages that are contained in the subject. Some of these examples are promising for future speculations. 

    Adaptive Multimedia Retrieval: From Data to User Interaction

    Well for this post i took a title from an article under the same name. On a side note, one of the contributors is Marcin Detyniecki who comes from Poland. What does this sentence mean actually?



    From what i understand, adaptive multimedia has an ability to dynamically suit user's needs. So, it's more than just and mp3 or video file that we can play back. It involves interaction with user perceiving the media with the media itself, and other users that might already or will find this media. In the level of interaction and perception, it comes close to augmented reality, when there is a two-way communication between the media and perceiver.


    Multimedia is a pretty old term despite the fact, that it was promoted as something new not far ago. One of the funny aspects was that few years ago many personal computers in Poland were marketed as "Multimedia", when in fact, in theoretical aspect, any PC with a screen, input device and PC buzzer are multimedia.


    But nowadays, multimedia systems progressed very much. The only sense that's still not very well implemented in those system is smelling sense, although, multimedia cinemas already offers a short movies that stimulates every human's sense, including smell. That cinemas right here are called "5D" i believe.


    Aside from stimulating senses, adaptive multimedia means, as it's self-explanatory, adapting to user's needs. That means dynamically calculating user's personal test and proposing other medias that might be interesting to him. That means the ability to leave a trace, in form of a comment, appendix, modifying, creating new media, to show other users our opinion, suggestion, to start a general discussion about the media. That of course also means that user can have an impact on the media, and the media will response back, creating mentioned earlier two-way communication.




    I will try to include examples of adaptive multimedia systems in next post, but let's just assume, it touches many many different aspects of human computation.

    poniedziałek, 21 listopada 2011

    Author presentation

    Frank Hopfgartner is post doctoral research fellow and member of the AI Group at the International Computer Science Institute (ICSI). He is of German descent and is currently residing at Berkeley University.



    His main research interest is on adaptive information retrieval, concept based retrieval and the automated evaluation of interactive information retrieval systems. Frank (co-) authored over 40 scientific publications in the field of multimedia retrieval and adaptive recommender systems. Further, he was involved in various other activities.


    I'm going to discuss his chapter of the book "Computation for Humanity—Information Technology to Advance Society " called "Multimedia Retrieval and Adaptive Systems". More on the subject in next posts, stay tuned.


    More about Frank and his topic of interests:
    http://www.icsi.berkeley.edu/~fh/


    Here's a paper talking about the subject in details:
    http://doras.dcu.ie/16264/


    This page lists Frank's publications throughout his science career:

    http://www.informatik.uni-trier.de/~ley/db/indices/a-tree/h/Hopfgartner:Frank.html



    And if you want to buy one of Frank's book, here's link to the online bookstore that offers it:
    http://www.bookcity.pl/A5325900/Understanding_Video_Retrieval/Frank_Hopfgartner