Would you be so kind and add error bars on your histograms? Either for the standard error of the mean or percentiles or confidence intervals. Boostrapping can make this very easy to implement.
I really don't understand why a programmer would put hard limits on stuff like this (besides limites thats comes from 2*32 for unsigned int in the source code). Anyway the UI for this specific feature would probably make it unuseable far before reaching any kind of limit.
Because the media industry primary goal is to produce videos that are not to be broadcasted on the web? Yeah television is the probably media of the future.
Also don't you think that the FOSS bashing policy by MS was also turning off a large amount of very skilled developers who likes to work with opensource tools and contribute to open source projects? Is this unintended consequence negligible compared to other corporate strategic areas such as sales and competition vs Linux servers?
The point of Tim Berners-Lee is to say that the vocabulary used to provide the nutritional information was standardized by FDA and related laws. ("defined by *terms* generated by the US Food and Drug Administration (FDA)", emphasis mine). The valuation of those properties or terms on this specific packaging are produced by Utz. To speak the semantic web lingua, the nutrition info ontology has been authored by FDA while the instance data on the package was authored by Utz reusing the FDA ontology.
So it is actually dangerous to market TOR to non tech savy people who do not systematically check that they are surfing only on https websites or orther encrypted protocols.
I guess you can harvest a great amount of passwords and other sensitive data by sniffing the http traffic of a single exit node.
I always wondered whether it is not possible to attack TOR with statistical analysis provided you can dedicate significant resources to it.
Suppose you are a big brother-style government agency with many computers and bandwith pipes dedicated to your goals. Could you not register a significant amounts of output and intermediate nodes (like say 10% of all nodes) that are specially improved to cooperatively log output HTTP traffic along with various web services session cookies, headers and originating IP addresses in a centralized DB and then use statistical analysis to identify the candidate source IP addresses of suspicious HTTP traffic?
First start by Norvig's book for a general overview of Machine Learning.
Then the best practical guide to implement backpropagation training for feed forward neural networks is by Le Cun and Bottou: http://leon.bottou.org/papers/lecun-98x (PDF or DjVu versions - 44 pages).
However backprop will only reach interesting convergence for 2 to 3 layers NN with labeled data as input which is not the type of architectures presented in On Intelligence.
To explore deep architectures such as the Hierarchical Temporal Memory introduced by Jeff Hawkins you should read recent papers on Deep Belief Networks by G. Hinton and Y. Bengio. They share interesting similarities with HTMs among which is the general architecture of the layered cortex as described by the mathematical models of the brain by Karl Friston.
DBNs however lacks the temporal / sequencial aspect of HTMs.
My personnal take is to use local predictive models such as 2-layers feed forward neural network trained using backprop to predict the future observed data and stack them into a deep structure similar to DBNs and HTMs.
The red5 opensource flash application server along with the openmeetings video conference / whiteboard application might help for organizing voice + video meetings. Clients just require a flash 10 plugin in their browser.
opeenmeetings allows for uploading/sharing office files with live preview using openoffice + pdf2swf and image files with imagemagick and also add a nice desktop sharing feature.
That's precisely how you can recognize the project is managed by an INRIA team.
I suppose they have written an AI to find project names by randomly sampling a word book and trying to make it related to the project topic. No human being can possibly achieve that.
For instance the Cell/BE processor allows C/C++ programmers to manage memory directly with the Memory Flow Controller to perform double-buffered asynchronous transfer of data between the main memory and the processor memory. Using Direct Memory Access, Cell users can achieve 98% of the peak performance on some applications: http://www-128.ibm.com/developerworks/power/library/pa-cellperf/ .
OSSEC HID is a very helpful simple tool to help protect your linux box (or most other OS). It watches the logs for you (ssh, apache, mail servers,...) and spot abnormal patterns registered in XML rules then send alert mail to the box admin and is able to blacklist the IP address of bruteforce attackers for some time to avoid being dosed or ssh-bruteforced.You can whitelist your common ip adresses to avoid being blacklisted by DOS attacks with forged IP packets.
It also maintains checksums for system files to help detect rootkits or other intrusion. For more details see the project page:
Slashdot Mirror
Would you be so kind and add error bars on your histograms? Either for the standard error of the mean or percentiles or confidence intervals. Boostrapping can make this very easy to implement.
I really don't understand why a programmer would put hard limits on stuff like this (besides limites thats comes from 2*32 for unsigned int in the source code). Anyway the UI for this specific feature would probably make it unuseable far before reaching any kind of limit.
Because the media industry primary goal is to produce videos that are not to be broadcasted on the web? Yeah television is the probably media of the future.
I love the new design. Thank you very much for your hard work.
Not with column store databases such as Cassandra, HBase and BigTable.
Also don't you think that the FOSS bashing policy by MS was also turning off a large amount of very skilled developers who likes to work with opensource tools and contribute to open source projects? Is this unintended consequence negligible compared to other corporate strategic areas such as sales and competition vs Linux servers?
The point of Tim Berners-Lee is to say that the vocabulary used to provide the nutritional information was standardized by FDA and related laws. ("defined by *terms* generated by the US Food and Drug Administration (FDA)", emphasis mine). The valuation of those properties or terms on this specific packaging are produced by Utz. To speak the semantic web lingua, the nutrition info ontology has been authored by FDA while the instance data on the package was authored by Utz reusing the FDA ontology.
With a little bit of more luck you will get one crash through the roof of your house and get the rain *into* your living room :)
or a zombie outbreak if you are out of luck.
An experimental (daily snapshot) version for ubuntu is available here: https://launchpad.net/~chromium-daily/+archive/ppa
So it is actually dangerous to market TOR to non tech savy people who do not systematically check that they are surfing only on https websites or orther encrypted protocols. I guess you can harvest a great amount of passwords and other sensitive data by sniffing the http traffic of a single exit node.
I always wondered whether it is not possible to attack TOR with statistical analysis provided you can dedicate significant resources to it. Suppose you are a big brother-style government agency with many computers and bandwith pipes dedicated to your goals. Could you not register a significant amounts of output and intermediate nodes (like say 10% of all nodes) that are specially improved to cooperatively log output HTTP traffic along with various web services session cookies, headers and originating IP addresses in a centralized DB and then use statistical analysis to identify the candidate source IP addresses of suspicious HTTP traffic?
First start by Norvig's book for a general overview of Machine Learning. Then the best practical guide to implement backpropagation training for feed forward neural networks is by Le Cun and Bottou: http://leon.bottou.org/papers/lecun-98x (PDF or DjVu versions - 44 pages). However backprop will only reach interesting convergence for 2 to 3 layers NN with labeled data as input which is not the type of architectures presented in On Intelligence. To explore deep architectures such as the Hierarchical Temporal Memory introduced by Jeff Hawkins you should read recent papers on Deep Belief Networks by G. Hinton and Y. Bengio. They share interesting similarities with HTMs among which is the general architecture of the layered cortex as described by the mathematical models of the brain by Karl Friston. DBNs however lacks the temporal / sequencial aspect of HTMs. My personnal take is to use local predictive models such as 2-layers feed forward neural network trained using backprop to predict the future observed data and stack them into a deep structure similar to DBNs and HTMs.
Boot time is very important for laptops when the sleep / hibernate feature is not well supported, prone to crash or slower than the boot it-self.
The red5 opensource flash application server along with the openmeetings video conference / whiteboard application might help for organizing voice + video meetings. Clients just require a flash 10 plugin in their browser. opeenmeetings allows for uploading/sharing office files with live preview using openoffice + pdf2swf and image files with imagemagick and also add a nice desktop sharing feature.
http://python.org/
That's precisely how you can recognize the project is managed by an INRIA team. I suppose they have written an AI to find project names by randomly sampling a word book and trying to make it related to the project topic. No human being can possibly achieve that.
For instance the Cell/BE processor allows C/C++ programmers to manage memory directly with the Memory Flow Controller to perform double-buffered asynchronous transfer of data between the main memory and the processor memory. Using Direct Memory Access, Cell users can achieve 98% of the peak performance on some applications: http://www-128.ibm.com/developerworks/power/library/pa-cellperf/ .
OSSEC HID is a very helpful simple tool to help protect your linux box (or most other OS). It watches the logs for you (ssh, apache, mail servers, ...) and spot abnormal patterns registered in XML rules then send alert mail to the box admin and is able to blacklist the IP address of bruteforce attackers for some time to avoid being dosed or ssh-bruteforced.You can whitelist your common ip adresses to avoid being blacklisted by DOS attacks with forged IP packets.
It also maintains checksums for system files to help detect rootkits or other intrusion. For more details see the project page:
http://www.ossec.net/main/
Unfortunately it is not yet packaged in all major linux distros so security updates will have to be applied manually.