Automated feature engineering for HTTP tunnel detection

Generating discriminative input features is a key requirement for achieving highly accurate classifiers. The process of generating features from raw data is known as feature engineering and it can take significant manual effort. In this paper we propose automated feature engineering to derive a suite of additional features from a given set of basic features with the aim of both improving classifier accuracy through discriminative features, and to assist data scientists through automation. Our implementation is specific to HTTP computer network traffic. To measure the effectiveness of our proposal, we compare the performance of a supervised machine learning classifier built with automated feature engineering versus one using human-guided features. The classifier addresses a problem in computer network security, namely the detection of HTTP tunnels. We use Bro to process network traffic into base features and then apply automated feature engineering to calculate a larger set of derived features. The derived features are calculated without favour to any base feature and include entropy, length and N-grams for all string features, and counts and averages over time for all numeric features. Feature selection is then used to find the most relevant subset of these features. Testing showed that both classifiers achieved a detection rate above 99.93% at a false positive rate below 0.01%. For our datasets, we conclude that automated feature engineering can provide the advantages of increasing classifier development speed and reducing development technical difficulties through the removal of manual feature engineering. These are achieved while also maintaining classification accuracy.

Associations of serum perfluoroalkyl acid levels with T-helper cell-specific cytokines in children: By gender and asthma status

Perfluoroalkyl acids (PFAAs) are a group of common chemicals that ubiquitously exist in wildlife and humans. Experimental data suggest that they may alter T-lymphocyte functioning in situ by preferentially enhancing the development of T-helper 2 (TH2)- and inhibiting TH1-lymphocyte development and might increase allergic inflammation, but few human studies have been conducted. To evaluate the association between serum PFAAs concentrations and T-lymphocyte-related immunological markers of asthma in children, and further to assess whether gender modified this association, 231 asthmatic children and 225 non-asthmatic control children from Northern Taiwan were recruited into the Genetic and Biomarker study for Childhood Asthma. Serum concentrations of ten PFAAs and levels of TH1 [interferon (IFN)-γ, interleukin (IL)-2] and TH2 (IL-4 and IL-5) cytokines were measured. The results showed that asthmatics had significantly higher serum PFAAs concentrations compared with the healthy controls. When stratified by gender, a greater number of significant associations between PFAAs and asthma outcomeswere found in males than in females. Among males, adjusted odds ratios for asthma among those with the highest versus lowest quartile of PFAAs exposure ranged from 2.59 (95% CI: 1.14, 5.87) for the perfluorobutanesulfonate (PFBS) to 4.38 (95% CI: 2.02, 9.50) for perfluorooctanesulfonate (PFOS); and serum PFAAs were associated positively with TH2 cytokines and inversely with TH1 cytokines among male asthmatics. Among females, no significant associations between PFAAs and TH2 cytokines could be detected. In conclusion, increased serum PFAAs levels may promote TH cell dysregulation and alter the availability of key TH1 and TH2 cytokines, ultimately contributing to the development of asthma that may differentially impact males to a greater degree than females. These results have potential relevance in asthma prevention.

The AOL Huffington Post merger

This article examines the merger between AOL and the Huffington Post. The broader issues around the merger will be investigated, especially the implication for rights, in particular free expression, and their conditions for exercise and actual exercise online. One major issue is that of the status of user-generated content and how the existing legal regime reflects the ethical concerns of users over how their content, data and information is used and commodified by the for-profit Internet intermediaries and platforms, especially when they start to merge and form concentrations. The extent to which the current legal regimes, especially human rights, deal with these problems in an adequate fashion will be assessed, along with the presentation of some suggestions of alternative approaches which may be more effective in practice.