hillarysanders deep_learning issues

hillarysanders / deep_learning

Learnin' about Deep Learnin!

1 stars 0 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

URL model comparisons

#30 hillarysanders opened 7 years ago
1
push iid dataset into an sqliite db

#29 hillarysanders closed 7 years ago
1
Upgrade to keras 2.0. Also, build a better infrastructure for testing models and saving states - e.g. each time a model is tested, the code to create that model, the code to create the features, should all be saved. And features, therefore, should be more modular.

#28 hillarysanders closed 7 years ago
1
read https://arxiv.org/pdf/1609.02226v3.pdf before april 13 for paper club

#27 hillarysanders opened 7 years ago
0
project layer 16 NN output to 2d, and plot (color by label) - show to josh

#26 hillarysanders closed 7 years ago
0
rf stuff: test on bagged logistic regression

#25 hillarysanders closed 7 years ago
1
what's stopping us from designing a neural net that cuts / grows connections between neurons over time? e.g. shuts down non-useful connections / neurons, and duplicates with noise neurons that are very useful...? Could that help the learning process?

#24 hillarysanders opened 7 years ago
0
design, built, and test out new NN for HTML detection in behavioral_model - need some code to do the cleaning

#23 hillarysanders closed 7 years ago
0
read more about NNs

#22 hillarysanders opened 7 years ago
0
clean up & submit blackhat proposal

#21 hillarysanders opened 7 years ago
1
read some Sophos whitepapers to get a better background on software security / malware detection in general and at Sophos

#20 hillarysanders opened 7 years ago
0
read / review some papers on review some papers on malicious HTML detection for next Thursday's talk

#19 hillarysanders closed 7 years ago
0
check that: master branch doesn't fit to iid any better than current

#18 hillarysanders closed 7 years ago
0
transform 2016-12-01_to_2017-02-24_by_0.1_days_x_10 to better csv --> db --> test rf on it

#17 hillarysanders closed 7 years ago
0
blackhat proposal ideas - read relevant papers and write up the 2 abstracts

#16 hillarysanders closed 7 years ago
0
clean up beahvioral_model code

#15 hillarysanders opened 7 years ago
1
investigate HTML VT data - possibility of building a malicious HTML detector

#14 hillarysanders closed 7 years ago
1
extract features from NN, use as input to RF

#13 hillarysanders closed 7 years ago
0
post chrome history script to slack channel, uploads to s3-bucket

#12 hillarysanders closed 7 years ago
2
sqlite --> CSV in convnet, data load

#11 hillarysanders closed 7 years ago
0
shut down small instance, run larger data fetch for plots? better iid

#10 hillarysanders closed 7 years ago
0
install tensorflow gpu on ubuntu instance

#9 hillarysanders closed 7 years ago
1
figure out how to run convnet on aws instance

#8 hillarysanders closed 7 years ago
0
fix RR!

#7 hillarysanders closed 7 years ago
0
write script to send e-mails, phone calls?

#6 hillarysanders opened 7 years ago
0
add parallel pooling capability to data fetch utils

#5 hillarysanders closed 7 years ago
0
make slides

#4 hillarysanders closed 7 years ago
0
make AUC by 5-1 plot

#3 hillarysanders closed 7 years ago
0
be able to replicate sqlite db for convnet fitting

#2 hillarysanders closed 7 years ago
0
look through keras code, create branch to extract features

#1 hillarysanders closed 7 years ago
0