A concurrent window system

This one is an oldy but goody. Rob Pike describes his implementation
of a windowing system using newsqueak.

A concurrent window system.

The system uses a well defined interface between the windowing system
and the client that supports simple synchronous messages (rather than
asynchronous events typical of windowing systems). This is one of the many
windowing systems Pike has written and you can see the precursors of some
Plan9 and 8-1/2 ideas.

Practical File System Design

This is a book which I have not read, but it looks interesting
enough that I wish I had, and I feel comfortable putting it
on paperchasin where others will find it interesting.

Practical file system design with the Be File System by D. Giampaolo.

Ghostbuster from M$

Not sure if you guys saw this... It came across TH-research list and I hadn't seen it before.

File hiding is an advanced stealth technique that is becoming popular among system monitoring software such as RootKits, Trojans, and keyloggers. It presents a major challenge to system administrators and the anti-malware industry because detection and removal are virtually impossible if the target files are not even visible. In this paper, we present the Strider GhostBuster that exploits the fundamental weakness of the file-hiding behavior and turns the problem into its own solution. We have tested this diff-based tool successfully in the lab against several real-world system monitoring programs. The simplicity and effectiveness of the approach suggest that the following quote on the Internet may no longer be true: “When you can get the dir command to lie, it’s all over.” In the post-GhostBuster world: “The best way to hide is not trying to hide.”
Keywords: Rootkit, Trojan, keylogger, spyware, Gatekeeper, WinPE, stealth, system monitoring software

Cheers.

http://research.microsoft.com/research/pubs/view.aspx?type=Technical%20Report&id=775

Co-Validation: Using Model Disagreement to Validate Classification Algorithms

I've been keenly interested in measuring classification techniques, this one popped up on Yahoo! research:

In the context of binary classification, we define disagreement as a measure of how often two independently-trained models differ in their classification of unlabeled data. We show that per-instance disagreement is an unbiased estimate of the variance of error for that instance. We also show that disagreement provides a lower bound on the prediction (generalization) error, and a tight upper bound on the “variance of prediction error”, or the variance of the average error (across instances), where variance is measured across training sets. We explore the use of disagreement for error estimation and model selection. We call the procedure co-validation, since the two models effectively (in)validate one another by comparing results on unlabeled data, which we assume is relatively cheap and plentiful compared to labeled data. The procedure is especially effective in active learning settings, where training sets are not drawn at random and cross validation overestimates error. We present experimental results on several data sets exploring co-validation for active learning and model selection.

Source: Co-Validation: Using Model Disagreement to Validate Classification Algorithms, Omid Madani, David M. Pennock, GaryW. Flake.

Scatter/Gather: A Cluster-based Approach to Browsing Large Document Collections

More data mining techniques, these authors utilize an interesting, and evidently efficient, mechanism to discover related items in a data corpus. It's described as such:

As an alternative, the Scatter/Gather interface uses text clustering as a way to group document according to the overall similarities in their content. Scatter/Gather is so named because it allows the user to scatter documents into clusters, or groups, then gather a subset of these groups and re-scatter them to form new groups.

Each cluster in Scatter/Gather is represented by a list of topical terms, that is, a list of words that attempt to give the user the gist of what the documents in the cluster are about. The user can also look at the titles of the documents in each group. The documents can in the cluster can have other representations as well, such as summaries, or TileBars.

If a cluster still has too many documents, the user can re-cluster the documents in the cluster; that is, re-group that subset of documents into still smaller groups. This re-grouping process tends to change the kinds of themes of the clusters, because the documents in a subcollection discuss a different set of topics than all the documents in the larger collection.

You can read more on their technique on the scatter/gather website. Unlike so many neat ideas in literature, this one is openly implemented. Papers and examples.

CS == SocialScience

I was looking for some unrelated information on Jon Pincus' website when I came across this interesting position paper. It states that computer science these days is mostly social science. It also has lots of references to interesting examples of cross polination.

http://research.microsoft.com/users/jpincus/cs%20SocSci.html


Also check out some of his other papers. Good stuff.

Entanglement Teleportation Through 1D Heisenberg Chain

This is relayed in from aempirei; lazy bastard ;-)

Entanglement Teleportation Through 1D Heisenberg Chain

In the words of aE:
"it made perfect sense now where all my socks have gone"

FreeBSD: UFS/FFS snapshot from high up

UFS/FFS flow

Pretty neat, while hard to see unless you print it out... we've all seen this kind of thing before, but figured this one was worth posting.

Basic Local Alignment Search Tool

This is one of the core papers in bioinformatics (dating from 1990), but has implications beyond the life sciences. Recall an earlier post of mine on string distance metrics.

A new approach to rapid sequence comparison, basic local alignment search tool (BLAST), directly approximates alignments that optimize a measure of local similarity, the maximal segment pair (MSP) score. Recent mathematical results on the stochastic properties of MSP scores allow analysis of the performance of this method as well as the statistical significance of alignments it generates. The basic algorithm is simple and robust; it can be implemented in a number of ways and applied in a variety of contexts including straight-forward DNA and protein sequence database searches, motif searches, gene identification searches, and in the analysis of multiple regions of similarity in long DNA sequences. In addition to its flexibility and tractability to mathematical analysis, BLAST is an order of magnitude faster than existing sequence comparison tools of comparable sensitivity.

Source: Basic Local Alignment Search Tool, Altschul, S.F., Gish, W., Miller, W., Meyers, E.W., Lipman, D.J.

An Introduction to Bayesian Networks and their Contemporary Applications

While so many people equate Bayesian techniques with Bayesian classifiers, ie for spam filtering, it has signficantly more applications than just spam filtering. This is one of the seminal papers on the topic of Bayesian Networks.

Bayesian Networks are becoming an increasingly important area for research and application in the entire field of Artificial Intelligence. This paper explores the nature and implications for Bayesian Networks beginning with an overview and comparison of inferential statistics and Bayes' Theorem. The nature, relevance and applicability of Bayesian Network theory for issues of advanced computability forms the core of the current discussion. A number of current applications using Bayesian networks is examined. The paper concludes with a brief discussion of the appropriateness and limitations of Bayesian Networks for human-computer interaction and automated learning.

Source: An Introduction to Bayesian Networks and their Contemporary Applications, Daryle Niedermayer.

Haystack: A Platform for Creating, Organizing and Visualizing Information Using RDF

I'm actually not a big fan of the "semantic web" or much of the XML community. I find it's too deeply smothered in politics, posturing, and wasted time. However, sometimes useful things come out of even idle dreaming, and one of the things I like is how people are tackling information management. The Haystack project is an attempt to implement a functional RDF browser on top of the Eclipse platform. While I don't use it regularily, I do find I'm intrigued by the techniques in information management I see being attempted.

The Resource Definition Framework (RDF) is designed to support agent communication on the Web, but it is also suitable as a framework for modeling and storing personal information. Haystack is a personalized information repository that employs RDF in this manner. This flexible semistructured data model is appealing for several reasons. First, RDF supports ontologies created by the user and tailored to the user’s needs. At the same time, system ontologies can be specified and evolved to support a variety of high-level functionalities such as flexible organization schemes, semantic querying, and collaboration. In addition, we show that RDF can be used to engineer a component architecture that gives rise to a semantically rich and uniform user interface. We demonstrate that by aggregating various types of users’ data together in a homogeneous representation, we create opportunities for agents to make more informed deductions in automating tasks for users. Finally, we discuss the implementation of an RDF information store and a programming language specifically suited for manipulating RDF.

Source: Haystack: A Platform for Creating, Organizing and Visualizing Information Using RDF, David Huynh, David Karger, and Dennis Quan, Semantic Web Workshop 2002 Hawaii, USA.

Forecasting Uncertain Events with Small Groups

... predicting future outcomes that use small numbers of individuals participating in an imperfect information market...

Forecasting Uncertain Events with Small Groups - K-Y. Chen, L. R. Fine, B. A. Huberman

The Baldwin Effect in the Immune System: Learning by Somatic Hypermutation

I read this paper a few years ago when doing some research on new methods or approaches to developing HIDS (not sure I was really thinking about NIDS at the time). However, it is an interesting medical paper that might be applicable to other fields; at least some theory regarding "learned or acquired characteristics" that "could become part of the genetic makeup of succeeding generations..."

The Baldwin Effect in the Immune System: Learning by Somatic Hypermutation - R. Hightower, S. Forrest, A. S. Perelson

Might as well add this paper too:

Myths and Legends of the Baldwin Effect - Peter Turney of Institute for Information Technology,
National Research Council Canada