Packet classification is a fundamental problem in computer networking. This problem exposes a hard tradeoff between the computation and state complexity, which makes it particularly challenging. To navigate this tradeoff, existing solutions rely on complex hand-tuned heuristics, which are brittle and hard to optimize.In this paper, we propose a deep reinforcement learning (RL) approach to solve the packet classification problem. There are several characteristics that make this problem a good fit for Deep RL. First, many existing solutions iteratively build a decision tree by splitting nodes in the tree. Second, the effects of these actions (e.g., splitting nodes) can only be evaluated once the entire tree is built. These two characteristics are naturally captured by the ability of RL to take actions that have sparse and delayed rewards. Third, it is computationally efficient to generate data traces and evaluate decision trees, which alleviate the notoriously high sample complexity problem of Deep RL algorithms. Our solution, NeuroCuts, uses succinct representations to encode state and action space, and efficiently explore candidate decision trees to optimize for a global objective. It produces compact decision trees optimized for a specific set of rules and a given performance metric, such as classification time, memory footprint, or a combination of the two. Evaluation on ClassBench shows that NeuroCuts outperforms existing hand-crafted algorithms in classification time by 18% at the median, and reduces both classification time and memory footprint by up to 3×.

No abstract

HCV can be present in the circulating blood either as a free virus or as a virion-immunoglobulin (Ig) complex. All isotypes of Igs may form the virus complexes, but it remains unclear what specific role of each Ig plays in the clearance of HCV. In the present study, we have combined immuno-capture and RT±PCR, and developed a quick double-specificity method for detecting and distinguishing different HCV-Ig complexes. We compared our new method, the immuno-capture RT±PCR (iRT±PCR), with the conventional RT±PCR (cRT±PCR) for the sensitivity of detecting HCV in 35 clinically diagnosed patients with HCV infection. The results showed that 31 patients were detected to be positive by using iRT±PCR, whereas 16 patients were positive with the use of cRT±PCR. HCV-IgM, HCV-IgG, HCV-IgA could separately be detected by iRT±PCR and their positive rates were 66.7%, 51.0%, 62.7%, respectively. HCV bound to antibody was a common feature of hepatitis C (HC) and 86.3% of patients were positive at least by one of the HCV-Ig tests. The patterns of HCV RNA constituents varied according to disease categories. In summary, iRT±PCR is a valuable method for analysis of the composition of the immune complexes, which may provide new and valuable insights into HCV pathogenesis.

No abstract