Yan Chen

Josh Abramson, Arun Ahuja, Federico Carnevale et al.

An important goal in artificial intelligence is to create agents that can both interact naturally with humans and learn from their feedback. Here we demonstrate how to use reinforcement learning from human feedback (RLHF) to improve upon simulated, embodied agents trained to a base level of competency with imitation learning. First, we collected data of humans interacting with agents in a simulated 3D world. We then asked annotators to record moments where they believed that agents either progressed toward or regressed from their human-instructed goal. Using this annotation data we leveraged a novel method - which we call "Inter-temporal Bradley-Terry" (IBT) modelling - to build a reward model that captures human judgments. Agents trained to optimise rewards delivered from IBT reward models improved with respect to all of our metrics, including subsequent human judgment during live interactions with agents. Altogether our results demonstrate how one can successfully leverage human judgments to improve agent behaviour, allowing us to use reinforcement learning in complex, embodied domains without programmatic reward functions. Videos of agent behaviour may be found at https://youtu.be/v_Z9F2_eKk4.

Zhenyuan Li, Jun Zeng, Yan Chen et al.

Cyber attacks are becoming more sophisticated and diverse, making detection increasingly challenging. To combat these attacks, security practitioners actively summarize and exchange their knowledge about attacks across organizations in the form of cyber threat intelligence (CTI) reports. However, as CTI reports written in natural language texts are not structured for automatic analysis, the report usage requires tedious manual efforts of cyber threat intelligence recovery. Additionally, individual reports typically cover only a limited aspect of attack patterns (techniques) and thus are insufficient to provide a comprehensive view of attacks with multiple variants. To take advantage of threat intelligence delivered by CTI reports, we propose AttacKG to automatically extract structured attack behavior graphs from CTI reports and identify the adopted attack techniques. We then aggregate cyber threat intelligence across reports to collect different aspects of techniques and enhance attack behavior graphs into technique knowledge graphs (TKGs). In our evaluation against 1,515 real-world CTI reports from diverse intelligence sources, AttacKG effectively identifies 28,262 attack techniques with 8,393 unique Indicators of Compromises (IoCs). To further verify the accuracy of AttacKG in extracting threat intelligence, we run AttacKG on 16 manually labeled CTI reports. Empirical results show that AttacKG accurately identifies attack-relevant entities, dependencies, and techniques with F1-scores of 0.887, 0.896, and 0.789, which outperforms the state-of-the-art approaches Extractor and TTPDrill. Moreover, the unique technique-level intelligence will directly benefit downstream security tasks that rely on technique specifications, e.g., APT detection and cyber attack reconstruction.

Ye Wang, Yan Chen, Haotian Wu et al.

Decentralized Exchanges (DEXes) enable users to create markets for exchanging any pair of cryptocurrencies. The direct exchange rate of two tokens may not match the cross-exchange rate in the market, and such price discrepancies open up arbitrage possibilities with trading through different cryptocurrencies cyclically. In this paper, we conduct a systematic investigation on cyclic arbitrages in DEXes. We propose a theoretical framework for studying cyclic arbitrage. With our framework, we analyze the profitability conditions and optimal trading strategies of cyclic transactions. We further examine exploitable arbitrage opportunities and the market size of cyclic arbitrages with transaction-level data of Uniswap V2. We find that traders have executed 292,606 cyclic arbitrages over eleven months and exploited more than 138 million USD in revenue. However, the revenue of the most profitable unexploited opportunity is persistently higher than 1 ETH (4,000 USD), which indicates that DEX markets may not be efficient enough. By analyzing how traders implement cyclic arbitrages, we find that traders can utilize smart contracts to issue atomic transactions and the atomic implementations could mitigate users' financial loss in cyclic arbitrage from the price impact.