Talks
Title:
Cryptography and Machine Learning: Past, Present and Future
Speaker:
Arpita Patra
Assistant Professor, Department of CSA, IISc Bangalore
Date:
26th June, 2019
Venue:
IIIT Bangalore
Abstract:
The interplay of cryptography and machine learning (ML) has reached a new height with the unprecedented possibility of using modern machine learning algorithms in diverse domains such as medical diagnosis, facial recognition, finance and many more. This talk will throw light on what future they hold together, what is the state-of-the-art and how they contributed each other in the past. Along the way, I will discuss where my cryptographic research fits in and can contribute to securing ML.
Biography of the speaker:
Arpita Patra is an Assistant Professor at Indian Institute of Science. Her area of interest is Cryptography, focusing on theoretical and practical aspects of secure multiparty computation protocols. She received her PhD from Indian Institute of Technology (IIT), Madras and held post-doctoral positions at University of Bristol, UK, ETH Zurich, Switzerland and Aarhus University, Denmark.
Title:
Fast Secure Computation for Small Population over the Internet
Speaker:
Arpita Patra
Assistant Professor, Department of CSA, IISc Bangalore
Date:
10th January, 2019
Venue:
IISc, Bangalore.
Abstract:
Secure Multi-Party Computation (MPC) with small number of parties is an interesting area of research, primarily due to its ability to model most real-life MPC applications and the simplicity and efficiency of the resulting protocols. We study constant-round 3-party (3PC) and 4-party (4PC) protocols in the honest-majority setting. On the theoretical front, we settle the exact round complexity of 3PC, for a range of security notions such as selective abort, unanimous abort, fairness and guaranteed output delivery. Selective abort security, the weakest in the lot, allows the corrupt parties to selectively deprive some of the honest parties of the output. In the mildly stronger version of unanimous abort, either all or none of the honest parties receive the output. Fairness implies that the corrupted parties receive their output only if all honest parties receive output and lastly, the strongest notion of guaranteed output delivery implies that the corrupted parties cannot prevent honest parties from receiving their output. On the practical front, we present efficient, constant-round 3PC and 4PC protocols that achieve strong security notions of fairness and guaranteed output delivery, with a nominal overhead compared to the known constructions achieving weaker notion of security. Being constant-round, our constructions are suitable for Internet-like high-latency networks and are built from garbled circuits (GC).
(Based on papers published in CRYPTO’18 with Divya Ravi and ACM CCS’18 with Megha Byali, Arun Joseph and Divya Ravi)
Biography of the speaker:
Arpita Patra is an Assistant Professor at Indian Institute of Science. Her area of interest is Cryptography, focusing on theoretical and practical aspects of secure multiparty computation protocols. She received her PhD from Indian Institute of Technology (IIT), Madras and held post-doctoral positions at University of Bristol, UK, ETH Zurich, Switzerland and Aarhus University, Denmark.
Title:
Fast Secure Computation for Small Population over the Internet
Speaker:
Arpita Patra
Assistant Professor, Department of CSA, IISc Bangalore
Date:
17th January, 2019
Time:
04:30 pm
Venue:
Technische Universität Darmstadt, Germany.
Abstract:
Secure Multi-Party Computation (MPC) with small number of parties is an interesting area of research, primarily due to its ability to model most real-life MPC applications and the simplicity and efficiency of the resulting protocols. We study constant-round 3-party (3PC) and 4-party (4PC) protocols in the honest-majority setting. On the theoretical front, we settle the exact round complexity of 3PC, for a range of security notions such as selective abort, unanimous abort, fairness and guaranteed output delivery. Selective abort security, the weakest in the lot, allows the corrupt parties to selectively deprive some of the honest parties of the output. In the mildly stronger version of unanimous abort, either all or none of the honest parties receive the output. Fairness implies that the corrupted parties receive their output only if all honest parties receive output and lastly, the strongest notion of guaranteed output delivery implies that the corrupted parties cannot prevent honest parties from receiving their output. On the practical front, we present efficient, constant-round 3PC and 4PC protocols that achieve strong security notions of fairness and guaranteed output delivery, with a nominal overhead compared to the known constructions achieving weaker notion of security. Being constant-round, our constructions are suitable for Internet-like high-latency networks and are built from garbled circuits (GC).
(Based on papers published in CRYPTO’18 with Divya Ravi and ACM CCS’18 with Megha Byali, Arun Joseph and Divya Ravi)
Biography of the speaker:
Arpita Patra is an Assistant Professor at Indian Institute of Science. Her area of interest is Cryptography, focusing on theoretical and practical aspects of secure multiparty computation protocols. She received her PhD from Indian Institute of Technology (IIT), Madras and held post-doctoral positions at University of Bristol, UK, ETH Zurich, Switzerland and Aarhus University, Denmark.
Title:
Cryptography and Machine Learning: Past, Present and Future
Speaker:
Arpita Patra
Assistant Professor, Department of CSA, IISc Bangalore
Date:
26th November, 2018
Venue:
LNM Institute of Information Technology, Jaipur.
Abstract:
The interplay of cryptography and machine learning (ML) has reached a new height with the unprecedented possibility of using modern machine learning algorithms in diverse domains such as medical diagnosis, facial recognition, finance and many more. This talk will throw light on what future they hold together, what is the state-of-the-art and how they contributed each other in the past. Along the way, I will discuss where my cryptographic research fits in and can contribute to securing ML.
Biography of the speaker:
Arpita Patra is an Assistant Professor at Indian Institute of Science. Her area of interest is Cryptography, focusing on theoretical and practical aspects of secure multiparty computation protocols. She received her PhD from Indian Institute of Technology (IIT), Madras and held post-doctoral positions at University of Bristol, UK, ETH Zurich, Switzerland and Aarhus University, Denmark.
Title:
Cryptography and Machine Learning: Past, Present and Future
Speaker:
Arpita Patra
Assistant Professor, Department of CSA, IISc Bangalore
Date:
27th October, 2018
Venue:
International Society of Automation, Bangalore.
Abstract:
The interplay of cryptography and machine learning (ML) has reached a new height with the unprecedented possibility of using modern machine learning algorithms in diverse domains such as medical diagnosis, facial recognition, finance and many more. This talk will throw light on what future they hold together, what is the state-of-the-art and how they contributed each other in the past. Along the way, I will discuss where my cryptographic research fits in and can contribute to securing ML.
Biography of the speaker:
Arpita Patra is an Assistant Professor at Indian Institute of Science. Her area of interest is Cryptography, focusing on theoretical and practical aspects of secure multiparty computation protocols. She received her PhD from Indian Institute of Technology (IIT), Madras and held post-doctoral positions at University of Bristol, UK, ETH Zurich, Switzerland and Aarhus University, Denmark.
Title:
Cryptography and Machine Learning: Past, Present and Future
Speaker:
Arpita Patra
Assistant Professor, Department of CSA, IISc Bangalore
Date:
12th October, 2018
Time:
04:00 pm
Venue:
CSA Seminar Hall (Room No. 254, First Floor)
Abstract:
The interplay of cryptography and machine learning (ML) has reached a new height with the unprecedented possibility of using modern machine learning algorithms in diverse domains such as medical diagnosis, facial recognition, finance and many more. This talk will throw light on what future they hold together, what is the state-of-the-art and how they contributed each other in the past. Along the way, I will discuss where my cryptographic research fits in and can contribute to securing ML.
Biography of the speaker:
Arpita Patra is an Assistant Professor at Indian Institute of Science. Her area of interest is Cryptography, focusing on theoretical and practical aspects of secure multiparty computation protocols. She received her PhD from Indian Institute of Technology (IIT), Madras and held post-doctoral positions at University of Bristol, UK, ETH Zurich, Switzerland and Aarhus University, Denmark.
Title:
Adaptively Secure Primitives in the Random Oracle Model
Speaker:
Pratik Sarkar
M.Tech (Research) Student, Department of CSA, IISc Bangalore
Date:
09th April, 2018
Time:
11:00 am
Abstract:
Adaptive security embodies one of the strongest notions of security that allows an adversary to corrupt parties at any point during protocol execution and gain access to its internal state. Since it models real-life situations such as ``hacking”, efficient adaptively-secure multiparty computation (MPC) protocols are desirable. Such protocols demand primitives such as zero knowledge (ZK), oblivious transfer (OT) and commitment schemes that are adaptively-secure as building blocks. Efficient realizations of these primitives have been found to be challenging, especially in the no erasure model. We make progress in this direction and provide efficient constructions that are Universally-Composable in the random oracle model.
Zero Knowledge: We construct an efficient UC-secure constant round ZK protocols from garbled circuits that are secure against adaptive corruptions, with communication linear in the size of the statement. Our work builds upon the efficient 5 round ZK protocol of Jawurek et al. (CCS 2013). We show that the ZK protocol can be made adaptively secure when the underlying oblivious transfer (OT) satisfies a mild adaptive security guarantee and a conditional verification technique is employed. The conditional verification technique gives us a three-round adaptively secure zero-knowledge argument in the plain random oracle model.
Oblivious Transfer: We present the first round optimal framework for building adaptively-secure OT in the programmable random oracle (PRO) model, relying upon the framework of Peikert et al. (Crypto 2008) that is only statically secure. When instantiated with Decisional Diffie Hellman assumption, it incurs a nominal communication overhead over its static counterpart. We complete the picture of efficient OT constructions by presenting the first adaptively secure OT Extension using PRO.
Commitment Scheme: We present an adaptively secure efficient commitment scheme solely relying on observable random oracle (ORO). Our commitment scheme has a one-time offline setup phase, where a common reference string (crs) is generated between the parties using an ORO. In the online phase, the parties use the crs and ORO to generate commitments in a non-interactive fashion.
