Thorpe, Christopher Andrew

Person:
Thorpe, Christopher Andrew

Last Name

Thorpe

First Name

Christopher Andrew

Name

Thorpe, Christopher Andrew

Full item page

Search Results

Now showing 1 - 7 of 7

Genetic Algorithm Optimization of Dynamic Support Vector Regression
(2009) Milnes, Thomas Bradford; Thorpe, Christopher Andrew; Pfeffer, Avi
We show that genetic algorithms (GA) find optimized dynamic support vector machines (DSVMs) more efficiently than the grid search (GS) optimization approach. In addition, we show that GA-DSVMs find extremely low-error solutions for a number of oft-cited benchmarks. Unlike standard support vector machines, DSVMs account for the fact that data further back in a time series are generally less predictive than more-recent data. In order to tune the discounting factors, DSVMs require two new free parameters for a total of five. Because of the five free parameters, traditional GS optimization becomes intractable for even modest grid resolutions. GA optimization finds better results while using fewer computational resources.
Intention-Disguised Algorithmic Trading
(2010) Yuen, William; Syverson, Paul; Zhenming, Liu; Thorpe, Christopher Andrew
We propose a general model underlying the problem of designing trading strategies that leak no information to frontrunners and other exploiters. We study major scenarios in the market and design a family of algorithms that can be proven to leak no information in important scenarios. These algorithms can serve as building blocks for more challenging real-world scenarios beyond our current scope. In contrast to previous work, the strategies we propose protect trader using the existing trading infrastructure.
Achieving Trust without Disclosure: Dark Pools and a Role for Secrecy-Preserving Verification
(2015) Parkes, David; Thorpe, Christopher Andrew; Li, Wei
Can an exchange be “dark,” so that orders are not displayed, while simultaneously trustworthy, so that the execution of trades and flow of information occur as promised? SEC actions against dark pools suggest cause for concern, and regulators seem to be moving towards requiring more disclosure. Yet there is a clear tension: trading order information is widely exploited. Therefore, institutional investors have a strong interest in keeping pre-trade information about large trades hidden. Secrecy-preserving proofs of correctness can be used to build trust without revealing unnecessary information. By performing operations on obfuscated representations of orders (perhaps encrypted or otherwise hidden), a zero knowledge proof can be provided, allowing anyone to verify correctness of trades. Crucially, this can be done without revealing any information beyond this correctness. This technology can be usefully applied to construct provably trustworthy dark pools. Additional practical protocols relax the definition of “zero knowledge" to reveal limited information, providing necessary transparency for efficient market operation while limiting information that can be exploited by observers. Coupled with Trusted Computing hardware, these protocols can provide an excellent balance of practicality with secrecy
Cryptographic Securities Exchanges
(Springer Nature, 2007) Thorpe, Christopher Andrew; Parkes, David
While transparency in financial markets should enhance liquidity, its exploitation by unethical and parasitic traders discourages others from fully embracing disclosure of their own information. Traders exploit both the private information in upstairs markets used to trade large orders outside traditional exchanges and the public information present in exchanges’ quoted limit order books. Using homomorphic cryptographic protocols, market designers can create “partially transparent” markets in which every matched trade is provably correct and only beneficial information is revealed. In a cryptographic securities exchange, market operators can hide information to prevent its exploitation, and still prove facts about the hidden information such as bid/ask spread or market depth.
Time-Lapse Cryptography
(2006) Rabin, Michael; Thorpe, Christopher Andrew
The notion of “sending a secret message to the future” has been around for over a decade. Despite this, no solution to this problem is in common use, or even attained widespread acceptance as a fundamental cryptographic primitive. We name, construct and specify an implementation for this new cryptographic primitive, “Time-Lapse Cryptography”, with which a sender can encrypt a message so that it is guaranteed to be revealed at an exact moment in the future, even if this revelation turns out to be undesirable to the sender. Our solution combines new ideas with Pedersen distributed key generation, Feldman verifiable threshold secret sharing, and ElGamal encryption, all of which rest upon the single, broadly accepted Decisional Diffie-Hellman assumption. We develop a Time-Lapse Cryptography Service (“the Service”) based on a network of parties who jointly perform the service. The protocol is practical and secure: at a given time T the Service publishes a public key so that anyone can use it, even anonymously. Senders encrypt their messages with this public key whose private key is not known to anyone – not even a trusted third party – until a predefined and specific future time T + δ, at which point the private key is constructed and published. At or after that time, anyone can decrypt the ciphertext using this private key. The Service is envisioned as a public utility publishing a continuous stream of encryption keys and subsequent corresponding time-lapse decryption keys. We complement our theoretical foundation with descriptions of specific attacks and defenses, and describe important applications of our service in sealed bid auctions, insider stock sales, clinical trials, and electronic voting.
Cryptographic Combinatorial Clock-Proxy Auctions
(Springer Verlag, 2009) Parkes, David; Rabin, Michael; Thorpe, Christopher Andrew
We present a cryptographic protocol for conducting efficient, provably correct and secrecy-preserving combinatorial clock-proxy auctions. The "clock phase" functions as a trusted auction despite price discovery: bidders submit encrypted bids, and prove for themselves that they meet activity rules, and can compute total demand and thus verify price increases without revealing any information about individual demands. In the sealed-bid "proxy phase", all bids are revealed the auctioneer via time-lapse cryptography and a branch-and-bound algorithm is used to solve the winner-determination problem. Homomorphic encryption is used to prove the correctness of the solution, and establishes the correctness of the solution to any interested party. Still an NP-hard optimization problem, the use of homomorphic encryption imposes additional computational time on winner-determination that is linear in the size of the branch-and-bound search tree, and thus roughly linear in the original (search-based) computational time. The result is a solution that avoids, in the usual case, the exponential complexity of previous cryptographically-secure combinatorial auctions.
Cryptographic Combinatorial Securities Exchanges
(Springer Verlag, 2009) Thorpe, Christopher Andrew; Parkes, David
We present a useful new mechanism that facilitates the atomic exchange of many large baskets of securities in a combinatorial exchange. Cryptography prevents information about the securities in the baskets from being exploited, enhancing trust. Our exchange offers institutions who wish to trade large positions a new alternative to existing methods of block trading: they can reduce transaction costs by taking advantage of other institutions' available liquidity, while third party liquidity providers guarantee execution--preserving their desired portfolio composition at all times. In our exchange, institutions submit encrypted orders which are crossed, leaving a "remainder". The exchange proves facts about the portfolio risk of this remainder to third party liquidity providers without revealing the securities in the remainder, the knowledge of which could also be exploited. The third parties learn either (depending on the setting) the portfolio risk parameters of the remainder itself, or how their own portfolio risk would change if they were to incorporate the remainder into a portfolio they submit. In one setting, these third parties submit bids on the commission, and the winner supplies necessary liquidity for the entire exchange to clear. This guaranteed clearing, coupled with external price discovery from the primary markets for the securities, sidesteps difficult combinatorial optimization problems. This latter method of proving how taking on the remainder would change risk parameters of one's own portfolio, without revealing the remainder's contents or its own risk parameters, is a useful protocol of independent interest.