State-of-the-art neural network models are actively used in various fields, but it is well-known that they are vulnerable to adversarial example attacks. Throughout the efforts to make the models robust against adversarial example attacks, it has been found to be a very difficult task. While many defense approaches were shown to be not effective, adversarial training remains as one of the promising methods. In adversarial training, the training data are augmented by “adversarial” samples generated using an attack algorithm. If the attacker uses a similar attack algorithm to generate adversarial examples, the adversarially trained network can be quite robust to the attack. However, there are numerous ways of creating adversarial examples, and the defender does not know what algorithm the attacker may use. A natural question is: Can we use adversarial training to train a model robust to multiple types of attack? Previous work have shown that, when a network is trained with adversarial examples generated from multiple attack methods, the network is still vulnerable to white-box attacks where the attacker has complete access to the model parameters. In this paper, we study this question in the context of black-box attacks, which can be a more realistic assumption for practical applications. Experiments with the MNIST dataset show that adversarially training a network with an attack method helps defending against that particular attack method, but has limited effect for other attack methods. In addition, even if the defender trains a network with multiple types of adversarial examples and the attacker attacks with one of the methods, the network could lose accuracy to the attack if the attacker uses a different data augmentation strategy on the target network. These results show that it is very difficult to make a robust network using adversarial training, even for black-box settings where the attacker has restricted information on the target network.
There are several automated stock trading programs using reinforcement learning, one of which is an ensemble strategy. The main idea of the ensemble strategy is to train DRL agents and make an ensemble with three different actor–critic algorithms: Advantage Actor–Critic (A2C), Deep Deterministic Policy Gradient (DDPG), and Proximal Policy Optimization (PPO). This novel idea was the concept mainly used in this paper. However, we did not stop there, but we refined the automated stock trading in two areas. First, we made another DRL-based ensemble and employed it as a new trading agent. We named it Remake Ensemble, and it combines not only A2C, DDPG, and PPO but also Actor–Critic using Kronecker-Factored Trust Region (ACKTR), Soft Actor–Critic (SAC), Twin Delayed DDPG (TD3), and Trust Region Policy Optimization (TRPO). Furthermore, we expanded the application domain of automated stock trading. Although the existing stock trading method treats only 30 Dow Jones stocks, ours handles KOSPI stocks, JPX stocks, and Dow Jones stocks. We conducted experiments with our modified automated stock trading system to validate its robustness in terms of cumulative return. Finally, we suggested some methods to gain relatively stable profits following the experiments.
Spatial reuse is an important factor in building highly efficient wireless local area networks. The key to maximizing spatial reuse is to support concurrent transmissions while avoiding packet losses due to interference. However, the current medium access control uses a conservative approach that discourages concurrent transmissions, and the recently introduced spatial reuse operation leads to a lot of packet collisions. This paper proposes two methods, CB-DCF (Contextual Bandit - Distributed Coordination Function) and CB-OBSS/PD (Contextual Bandit - Overlapping Basic Service Set / Preamble Detection), that can improve throughput and fairness of the network. In the proposed methods, nodes run contextual bandit agents that learn to transmit or defer under given states. In CB-DCF, a state consists of the packet destination, the interferer, and the interferer’s destination. Nodes periodically exchange Q tables to prevent concurrent transmissions from destroying the ongoing packet. In CB-OBSS/PD, instead of exchanging Q tables, a node reduces its transmit power when concurrently transmitting with the ongoing packet. While CB-OBSS/PD is simpler to implement, performance of CB-DCF is superior to CB-OBSS/PD. Through extensive simulations with the ns-3 simulator, we show that the proposed methods can achieve high throughput and fairness by intelligently controlling transmissions under various interference conditions.
With the density of wireless networks increasing rapidly, one of the major goals in next-generation wireless LANs (Local Area Networks) is to support a very dense network with a large number of closely deployed APs (Access Points) and crowded users. However, the CSMA (Carrier-Sense Multiple Access)-based medium access control of current wireless network systems suffers from significantly degraded performance when the network becomes dense. Recent WLAN (Wireless Local Area Networks) standards include measures for increasing spatial reuse such as BSS (Basic Service Set) coloring, but the schemes based on BSS coloring such as OBSS/PD (Overlapping BSS/Preamble Detection) have limitations in improving spatial reuse. In this paper, we propose a spatial reuse method for uplink which can utilize BSS color and proximity information to improve the efficiency of carrier sensing and thus spatial reuse. Specifically, through the BSS color and the proximity information, a node receiving a preamble can figure out how far the receiver of the ongoing traffic is located. This information is used to determine whether the node should aggressively start transmitting or defer its transmission to protect the ongoing transmission. Simulation results show that the proposed method outperforms existing methods in terms of throughput and fairness.