K. Lund, C. Burgess, and C. Audet. Dissociating semantic and associative word relationships using high-dimensional smeantic space. In Proceeding of the Cognitive Science Society, pages 603–608, Hillsdale, N.J., USA, 1996. Erlbaum Press. [pdf]

————————

The paper studies the lexical/semantic priming effect which is questioned to be associative in nature. The aim of the paper is to shed some light on the question for which two crucial point are tackled: firstly, an operational definition of of semantic and association is needed; secondly, the definition of a framework for modelling semantic representations.

Their proposition for the first point is that semantically related words (TABLE – BED) are instances of the same category and share a number of features. Associated words (MOLD – BREAD) are those which are associated as determined by human word association norms. There is also a third type that are both semantically and associatively related (UNCLE – AUNT).

To solve the second point they propose a framework, called HAL (Hyperspace Analogue to Language) that allow to simulate different experiments. The methodology is based on the computation of a matrix of co-occurrence vectors for each word, which can be analyzed for semantic content. The co-occurrence is defined using the window-size parameter (co-occurrence within n words). Than a similarity is computed between the vectors using an Eucledian distance measure.

Using a certain dataset, the author simulated a certain kind of association between word pairs. They they repeated the experiment with human subject and confronted the results. The conclusion was that the notion of associativity can be characterized by temporal association in language receive little or no support from their corpus analysis. Word association seeed to be more a function of semantic neighborhood.

Another interesting result was that the distinction between associative and semantic information corresponds to the distinction between local co-occurrence and global co-occurrence. Temporal information is reflected in local co-occurrence. Global pattern of co-occurrence across a vocabulary is connected to semantic information.

Tags: information retrieval

I found this nice application that allows to syncronise bookmarks, calendars, contacts, keychain, and mail between two or more Mac without buying an annoying .Mac account. Once MySync is installed on each computer to be synced, one copy of MySync needs to be configured as a “Master” node. The remaining copies should be configured as “Slave” nodes.

Once configured, the MySync nodes will automatically discover each other using Bonjour.

You will also need to enable syncing for the datatypes that you want to sync, and to run your first Sync manually; Apple’s Sync Framework will present a dialog on the first sync, asking you to confirm that MySync should be allowed to sync the specified data types. After your first sync, you can configure each MySync Slave node to sync automatically on a quarter hourly, hourly, daily, or weekly basis, or leave syncing to be activated manually.

Tags: hack, mac

Komodo seems to offer support for the native version of Python installed on OS X. Lots of nice features and the pricing for personal students use seems pretty decent. However there is no SVN support on the student version.

Tags: Python

Continue reading →

This workshop will focus on learning about the tools available to build these new kinds of spatial applications, and reflect on thoughts from philosophy, urban planning, computer science, design, and geographic information systems. Students in the workshop will be expected to attend lectures and framing sessions where we will go over tools, technologies, and techniques for building spatial applications. There will be friday critique sessions where students will have the opportunity to share their progress, ask questions, and receive feedback.

The workshop will culminate in a final design competition where the most provocative, interesting, and successful projects will receive recognition and where the “best of these” as determined by a panel of judges will be considered for the round and flat awards.

Additional Information:

spatialworkshop@media.mit.edu

M. Sahlgren. An introduction to random indexing. In Proceeding of the Workshop at the 7th International Conference on Terminology and Knowledge Engineering, conference Methods and Applications of Semantic Indexing, Copenhagen, Denmark, 16th August 2005. [pdf]

———————–

This paper presents the Random Indexing algoritm that is introduced as a good alternative to LSI and similar word space methods that are based on the distributional hypothesis, which states that words with similar meaning tend to occur in similar contexts.

The author states the limit of word space models, which is the efficiency and the scalability problem: the co-occurrence matrix will become soon computationally intractable when the vocabulary grows. Additionally, the author highlight the fact that a majority of the cells will be zero: a tiny amount of the words in language are distributionally promiscuous; the vast majority of words only occur in a very limited set of contexts.

Available methods, like LSI  uses for this purpose a matrix truncation to reduce the dimensionality that according to the author should be avoided for three reasons: 1) the reduction is computationally costly; 2) the reduction is one-time operation that needs to be redone each time a new dimension is added; 3) the reduction requires initial sampling of the data that is often done with ad-hoc solutions.

Random Indexing goes on the contrary of traditional views (i.e., first construct co-occurrence matrix, then extract context vectors) moving from the accumulation of context vecotrs to the construction of the co-occurrence matrix.

Tags: information retrieval, Latent Semantic Analysis, Singular Value Decomposition

This is the first post of the new year, which both should summarize the ended year 2005 and set the resolutions for the coming time. First of all I should say that 2005 was a tremendous year for blogging. The thing just boomed see this personal stat:

• 2002 -> 7 posts
• 2003 -> 116 posts
• 2004 -> 333 posts (almost one per day)
• 2005 -> 539 posts (almost two per day)

Using this tool I achieved two great objectives: first, to have a log of the evolution of my research subject and conclusion; second, the construction of the network of people working on the same subject (I have got 119 comments so far, out of which a dozen were real and good research contacts). A third objective was that of communicating my updated research status to my co-workers that regularly check my blog, and finally the good point was to push me to be less lazy.

I can state with no doubts that blogging and aggregation are now a fundamental tool for research and learning in general. This is obviously a revolution in the internet revolution: the shift of control on content production. Internet 1.0 was a huge showcase. Internet 2.0 is a constantly evolving people-based environment, where ‘fruitors’ are in control of the content.

Some great scenarios are just one simple step ahead: the redefinition of streamline media power of information into a citizen-driven domain. For the new year I would love to see less posts but with better and more original content. What I would love is to reblog less and talk more in detail of my own points of view. May the force be with us in this 2006! Happy blogging!!

Tags: blogs, citizen journalism, collaboration tools, learning technology

A great resource for classroom teaching and lifelong learning: a freely downloadable physics textbook that spans over the fundamentals of the field with an eye on the recent research:

The project aims to produce a simple, vivid and up-to-date introduction to modern physics, with emphasis on the fundamental ideas of motion. ‘Simple’ means that concepts are stressed more than formalism; ‘vivid’ means that the reader is continuously challenged; ‘up-to-date’ means that modern research and ideas about unification are included.

The project takes the search for a precise description of motion as a guiding principle for an exploration of modern physics. This leads to an approach which is somewhat different from the standard one. Nature’s limits to speed, entropy, force, action and charge are central to the presentation.

Other physics textbooks online:

The high-quality electrodynamics text by Bo Thidé, Electromagnetic Field Theory, can be found at http://www.plasma.uu.se/CED/Book, together with its collection of exercises.

There are also the excellent introductory physics texts by Benjamin Crowell, Simple Nature and Light and Matter, which can be downloaded at http://www.lightandmatter.com.

The interesting Biophysics Educational Resources, edited by Victor Bloomfield, can be found at http://www.biophysics.org/education/resources.htm.

The text Classical and Quantum Chaos, edited by Predrag Cvitanovic, Roberto Artuso, Per Dahlqvist, Ronnie Mainieri, Gregor Tanner, Gábor Vattay, Niall Whelan and Andreas Wirzba can be found at http://chaosbook.org.

Tags: learning technology, new technology, open source, pedagogy, politics

Some recent comments on my blog captured my attention on this interesting (and refreshing, to speak as Nicolas) project on LBSs. From the web site description I inferred that the team involved is interested more on the development of the infrastructure that specific services.

The research in this area is based on a scenario where consumers are offered services in a transparent fashion as they move between environments with heterogeneous communication infrastructures. The goal is to provide the users with ubiquitous access to an open and expanding set of affordable services. In this research project, we will focus on professional collaborative work environments as well as high quality audio/visual entertainment applications. This involves the utilization of IT artifacts (devices, terminals) that dissolve into small components that increasingly melt into the background and into the fabric of everyday things, “smart dust” being an extreme case. Ad-hoc functionality on all systems levels implies dynamic and automatic configuration of public and private artifacts and services. Both the technical solutions as well as possible business models are of importance.

Tags: location awareness, Location Based Services

The MOGA-game has been developed as a part of the Rotuaari research project in the University of Oulu. MOGA’s goal is to test how WiFi positioning works inside campus buildings and are end users willing to move from place to place during actual game play.

MOGA’s idea is an old one: capture the flag. The game is played in teams. The idea is to capture and defend virtual flag zones. Game is played with a mobile handheld device, in which player’s position and the flag zones are shown on the screen. Cooperation within a team is emphasized: players are forced to play as a team. Capturing flags is much faster, when it’s done as a team, but also flags situated nearby have effects. Players must also pay attention to the flags positions in their strategy and their affect to the game itself.

MOGA has been implemented but testing is still in progress. MOGA is a part of Heikki Tolonen’s Masters’ Thesis that concentrates in heuristic evaluation of location-aware mobile games (Heikki contact’s: heikki.tolonen (at) @ee.oulu.fi).

Tags: location awareness, Location Based Services, ubiquitous game, urban exploration

Hughes Research has developed this interesting idea: a ramp for cars that converts some of the car movement energy in electrical energy. Basically when a car pass over the ramp, the ramp get pressed, activating an alternator that charge an accumulator.

I don’t think is a revolutionary concept, as no new energy is produced, but in essence can be a powerful way to tax the car drivers of some energetical value.

Tags: ecology, environment, new technology, politics