Speech Train + Proxy Server and all

Oops, I accidentally enable proxy in my apache server (need a proxy so I can cache the fedora rpm files when building fedora livecd for speech train machines) and it was public on the internet on port 80. Well, you can image what happened. Loads of people use it;). My router constantly went down for unknown reason. I figured out the problem by disconnecting ethernet cable from the router one by one.

Surprisingly, most proxied request was not porn;). It's Ads. Almost 60% are ads. Briefly looked through the log file, most connections were from Russian and Japan. A few porn sites were logged there.

Darn, had to disable the proxy. Sorry, folks.

Speech Training and all

after 3 weeks og tweaterring around, I finally get 1500 hours of speech ready to feed into sphinx. My estimate for the training is 1 month (at least). no more athena help... Hmm, therotically, I can log into a few athena machines and use it in my cluster, but I'd have to stay in front of the machine all the time.

And, distribute on the internet wouldn't work, given the huge data size. oh, well, we shall see.

Job Hunt and all

Jobs, jobs, jobs...

I feel I just got in yesterday, but now it's time for jobs. What an experience. I do think I have done a lot though. Really enjoyed it here, but not sure the reasons. Maybe the people.

So many thing I have learned. The entrepreneurship thing, I am enjoying it. I wouldn't say the team is ideal, but hey, there is no such as ideal. Everyone is different and you have to work with what you got.

oh, I should also connect with my classmates more often, like actually go to the classes;). I also should balance study and work more. Oh, and job hunt.

Back to jobs, I am facing a little dilemma: if I go to those off-campus job hunt site, like monster.com, the jobs there are most likely immediate openings. The jobs available through campus career fair are mostly geared towards campus new-hire with no professional experience. Oh, well, I do have a few leads though.

Warning from Athena

darn, Athena admin obviously wasn't too happy of my recent athena activity. I received a wraning note from athena admin. or, whatever, I'll stick with my own cluster then. well, maybe I'll have short memory...

Benjamin,

Most recently our team received a report of configuration changes made to several hosts in the W20-575 Athena cluster on September 7 in the early morning. Your computing account was logged into console at the time these changes were made. Specifically, the TORQUE resource manager was installed and several workstations were reconfigured as compute nodes. The head node (hourglass.mit.edu) also appears to be a host belonging to you as well.

Reconfiguring Athena workstations is a serious violation of the Athena Rules of Use http://mit.edu/olh/Rules/), because it interferes with other people's access to these resources. Athena is a shared resource and should be equally available to all members of the Athena community.

All Athena account holders are responsible for the actions from their account, and repeated violations of the Rules can lead to account suspension and termination. We expect that you will reacquaint yourself with your responsibilities under the Rules and abide by them.

Regards,

Brass Rat is here

After 2 month of waiting, THE Brass Rat is finally here. It is ... visible;).

Ouch, I haven't updated my blog for almost two months now. What a busy summer. The venture idea is taking up a lot of time. Hopefully, I can have more free time soon.


Ben

Friday Accounting and Beer...

The only good thing this summer is that Linda from 07 organized the Beer thing after Friday accounting class. It's good to meet SDMer at muddy.

Love it and thanks Linda.

What have I been doing....

Oh, well, long time no post. Let me check what I have done:

1. trying to start our business (or thesis, however we want to call it) with Cynthia and the rest of the gang, a lot of coding in the 3 weeks break.

2. Summer session has officially started. Classes seem really hectic now. Homework piles on and I really want to get the coding part to a certain milestone.

3. doing a lot of reading and experiment with Automated Speech Recognition (ASR). Very interesting field. Right now, doing some model training with Sphinx. It is so much .... not-fun. Well, a lot of things need to be flushed out.

Something interesting in the process was that we (Cynthia and I) had a visit to Virage about 2 weeks ago. Their VP of research gave us some good demo about their product, eViTAP and other (blinkx is one of their product). Amazing speech recognition toolkit and machine translation capbility.

Other than these, I will remember to write more.