Defense against Joint Poisoning and Evasion Attacks: A Case Study of DERMS. ul Abdeen, Z., Roy, P., Al-Tawaha, A., Jia, R., Freeman, L., Beling, P., Liu, C., Sangiovanni-Vincentelli, A., & Jin, M. Preprint, 2022. ![Defense against Joint Poisoning and Evasion Attacks: A Case Study of DERMS [pdf]](https://bibbase.org/img/filetypes/pdf.svg "pdf")
Pdf abstract bibtex 27 downloads There is an upward trend of deploying distributed energy resource management systems (DERMS) to control modern power grids. However, DERMS controller communication lines are vulnerable to cyberattacks that could potentially impact operational reliability. While a data-driven intrusion detection system (IDS) can potentially thwart attacks during deployment, also known as the evasion attack, the training of the detection algorithm may be corrupted by adversarial data injected into the database, also known as the poisoning attack. In this paper, we propose the \emphfirst framework of IDS that is robust against joint poisoning and evasion attacks. We formulate the defense mechanism as a bilevel optimization, where the inner and outer levels deal with attacks that occur during training time and testing time, respectively. We verify the robustness of our method on the IEEE-13 bus feeder model against a diverse set of poisoning and evasion attack scenarios. The results indicate that our proposed method outperforms the baseline technique in terms of accuracy, precision, and recall for intrusion detection.
@article{2022_2C_CyberPower,
title={Defense against Joint Poisoning and Evasion Attacks: A Case Study of DERMS},
author={Zain ul Abdeen and Padmaksha Roy and Ahmad Al-Tawaha and Rouxi Jia and Laura Freeman and Peter Beling and Chen-Ching Liu and Alberto Sangiovanni-Vincentelli and Ming Jin},
year={2022},
journal = {Preprint},
url_pdf={Cybersecurity_PS2022.pdf},
keywords = {Machine Learning, Cybersecurity, Power system},
abstract={There is an upward trend of deploying distributed energy resource management systems (DERMS) to control modern power grids. However, DERMS controller communication lines are vulnerable to cyberattacks that could potentially impact operational reliability. While a data-driven intrusion detection system (IDS) can potentially thwart attacks during deployment, also known as the evasion attack, the training of the detection algorithm may be corrupted by adversarial data injected into the database, also known as the poisoning attack. In this paper, we propose the \emph{first} framework of IDS that is robust against joint poisoning and evasion attacks. We formulate the defense mechanism as a bilevel optimization, where the inner and outer levels deal with attacks that occur during training time and testing time, respectively. We verify the robustness of our method on the IEEE-13 bus feeder model against a diverse set of poisoning and evasion attack scenarios. The results indicate that our proposed method outperforms the baseline technique in terms of accuracy, precision, and recall for intrusion detection. }
}
Downloads: 27
{"_id":"v2wvZasEdsuALnubb","bibbaseid":"ulabdeen-roy-altawaha-jia-freeman-beling-liu-sangiovannivincentelli-etal-defenseagainstjointpoisoningandevasionattacksacasestudyofderms-2022","author_short":["ul Abdeen, Z.","Roy, P.","Al-Tawaha, A.","Jia, R.","Freeman, L.","Beling, P.","Liu, C.","Sangiovanni-Vincentelli, A.","Jin, M."],"bibdata":{"bibtype":"article","type":"article","title":"Defense against Joint Poisoning and Evasion Attacks: A Case Study of DERMS","author":[{"firstnames":["Zain"],"propositions":["ul"],"lastnames":["Abdeen"],"suffixes":[]},{"firstnames":["Padmaksha"],"propositions":[],"lastnames":["Roy"],"suffixes":[]},{"firstnames":["Ahmad"],"propositions":[],"lastnames":["Al-Tawaha"],"suffixes":[]},{"firstnames":["Rouxi"],"propositions":[],"lastnames":["Jia"],"suffixes":[]},{"firstnames":["Laura"],"propositions":[],"lastnames":["Freeman"],"suffixes":[]},{"firstnames":["Peter"],"propositions":[],"lastnames":["Beling"],"suffixes":[]},{"firstnames":["Chen-Ching"],"propositions":[],"lastnames":["Liu"],"suffixes":[]},{"firstnames":["Alberto"],"propositions":[],"lastnames":["Sangiovanni-Vincentelli"],"suffixes":[]},{"firstnames":["Ming"],"propositions":[],"lastnames":["Jin"],"suffixes":[]}],"year":"2022","journal":"Preprint","url_pdf":"Cybersecurity_PS2022.pdf","keywords":"Machine Learning, Cybersecurity, Power system","abstract":"There is an upward trend of deploying distributed energy resource management systems (DERMS) to control modern power grids. However, DERMS controller communication lines are vulnerable to cyberattacks that could potentially impact operational reliability. While a data-driven intrusion detection system (IDS) can potentially thwart attacks during deployment, also known as the evasion attack, the training of the detection algorithm may be corrupted by adversarial data injected into the database, also known as the poisoning attack. In this paper, we propose the \\emphfirst framework of IDS that is robust against joint poisoning and evasion attacks. We formulate the defense mechanism as a bilevel optimization, where the inner and outer levels deal with attacks that occur during training time and testing time, respectively. We verify the robustness of our method on the IEEE-13 bus feeder model against a diverse set of poisoning and evasion attack scenarios. The results indicate that our proposed method outperforms the baseline technique in terms of accuracy, precision, and recall for intrusion detection. ","bibtex":"@article{2022_2C_CyberPower,\n title={Defense against Joint Poisoning and Evasion Attacks: A Case Study of DERMS},\n author={Zain ul Abdeen and Padmaksha Roy and Ahmad Al-Tawaha and Rouxi Jia and Laura Freeman and Peter Beling and Chen-Ching Liu and Alberto Sangiovanni-Vincentelli and Ming Jin},\n year={2022},\n journal = {Preprint}, \n url_pdf={Cybersecurity_PS2022.pdf},\n keywords = {Machine Learning, Cybersecurity, Power system},\n abstract={There is an upward trend of deploying distributed energy resource management systems (DERMS) to control modern power grids. However, DERMS controller communication lines are vulnerable to cyberattacks that could potentially impact operational reliability. While a data-driven intrusion detection system (IDS) can potentially thwart attacks during deployment, also known as the evasion attack, the training of the detection algorithm may be corrupted by adversarial data injected into the database, also known as the poisoning attack. In this paper, we propose the \\emph{first} framework of IDS that is robust against joint poisoning and evasion attacks. We formulate the defense mechanism as a bilevel optimization, where the inner and outer levels deal with attacks that occur during training time and testing time, respectively. We verify the robustness of our method on the IEEE-13 bus feeder model against a diverse set of poisoning and evasion attack scenarios. The results indicate that our proposed method outperforms the baseline technique in terms of accuracy, precision, and recall for intrusion detection. }\n}\n\n","author_short":["ul Abdeen, Z.","Roy, P.","Al-Tawaha, A.","Jia, R.","Freeman, L.","Beling, P.","Liu, C.","Sangiovanni-Vincentelli, A.","Jin, M."],"key":"2022_2C_CyberPower","id":"2022_2C_CyberPower","bibbaseid":"ulabdeen-roy-altawaha-jia-freeman-beling-liu-sangiovannivincentelli-etal-defenseagainstjointpoisoningandevasionattacksacasestudyofderms-2022","role":"author","urls":{" pdf":"http://www.jinming.tech/papers/Cybersecurity_PS2022.pdf"},"keyword":["Machine Learning","Cybersecurity","Power system"],"metadata":{"authorlinks":{}},"downloads":27},"bibtype":"article","biburl":"http://www.jinming.tech/papers/myref.bib","dataSources":["sTzDHHaipTZWjp8oe","Y64tp2HnDCfXgLdc5"],"keywords":["machine learning","cybersecurity","power system"],"search_terms":["defense","against","joint","poisoning","evasion","attacks","case","study","derms","ul abdeen","roy","al-tawaha","jia","freeman","beling","liu","sangiovanni-vincentelli","jin"],"title":"Defense against Joint Poisoning and Evasion Attacks: A Case Study of DERMS","year":2022,"downloads":27}