-
Exploring the Generalization of Cancer Clinical Trial Eligibility Classifiers Across Diseases
Authors:
Yumeng Yang,
Ashley Gilliam,
Ethan B Ludmir,
Kirk Roberts
Abstract:
Clinical trials are pivotal in medical research, and NLP can enhance their success, with application in recruitment. This study aims to evaluate the generalizability of eligibility classification across a broad spectrum of clinical trials. Starting with phase 3 cancer trials, annotated with seven eligibility exclusions, then to determine how well models can generalize to non-cancer and non-phase 3…
▽ More
Clinical trials are pivotal in medical research, and NLP can enhance their success, with application in recruitment. This study aims to evaluate the generalizability of eligibility classification across a broad spectrum of clinical trials. Starting with phase 3 cancer trials, annotated with seven eligibility exclusions, then to determine how well models can generalize to non-cancer and non-phase 3 trials. To assess this, we have compiled eligibility criteria data for five types of trials: (1) additional phase 3 cancer trials, (2) phase 1 and 2 cancer trials, (3) heart disease trials, (4) type 2 diabetes trials, and (5) observational trials for any disease, comprising 2,490 annotated eligibility criteria across seven exclusion types. Our results show that models trained on the extensive cancer dataset can effectively handle criteria commonly found in non-cancer trials, such as autoimmune diseases. However, they struggle with criteria disproportionately prevalent in cancer trials, like prior malignancy. We also experiment with few-shot learning, demonstrating that a limited number of disease-specific examples can partially overcome this performance gap. We are releasing this new dataset of annotated eligibility statements to promote the development of cross-disease generalization in clinical trial classification.
△ Less
Submitted 25 March, 2024;
originally announced March 2024.
-
Transfers between moons with escape and capture patterns via Lyapunov exponent maps
Authors:
David Canales,
Kathleen C. Howell,
Elena Fantino,
Annika J. Gilliam
Abstract:
This contribution focuses on the design of low-energy transfers between planetary moons and presents an efficient technique to compute trajectories characterized by desirable behaviors in the vicinities of the departure and destination bodies. The method utilizes finite-time Lyapunov exponent maps in combination with the Moon-to-Moon Analytical Transfer (MMAT) method previously proposed by the aut…
▽ More
This contribution focuses on the design of low-energy transfers between planetary moons and presents an efficient technique to compute trajectories characterized by desirable behaviors in the vicinities of the departure and destination bodies. The method utilizes finite-time Lyapunov exponent maps in combination with the Moon-to-Moon Analytical Transfer (MMAT) method previously proposed by the authors. The integration of these two components facilitates the design of direct transfers between moons within the context of the circular restricted three-body problem, and allows the inclusion of a variety of trajectory patterns, such as captures, landings, transits and takeoffs, at the two ends of a transfer. The foundations and properties of the technique are illustrated through an application based on impulsive direct transfers between Ganymede and Europa. However, the methodology can be employed to assist in the design of more complex mission scenarios, such as moon tours.
△ Less
Submitted 19 August, 2023;
originally announced August 2023.
-
Optimizing Quantum Search with a Binomial Version of Grover's Algorithm
Authors:
Austin Gilliam,
Marco Pistoia,
Constantin Gonciulea
Abstract:
Amplitude Amplification -- a key component of Grover's Search algorithm -- uses an iterative approach to systematically increase the probability of one or multiple target states. We present novel strategies to enhance the amplification procedure by partitioning the states into classes, whose probabilities are increased at different levels before or during amplification. The partitioning process is…
▽ More
Amplitude Amplification -- a key component of Grover's Search algorithm -- uses an iterative approach to systematically increase the probability of one or multiple target states. We present novel strategies to enhance the amplification procedure by partitioning the states into classes, whose probabilities are increased at different levels before or during amplification. The partitioning process is based on the binomial distribution. If the classes to which the search target states belong are known in advance, the number of iterations in the Amplitude Amplification algorithm can be drastically reduced compared to the standard version. In the more likely case in which the relevant classes are not known in advance, their selection can be configured at run time, or a random approach can be employed, similar to classical algorithms such as binary search. In particular, we apply this method in the context of our previously introduced Quantum Dictionary pattern, where keys and values are encoded in two separate registers, and the value-encoding method is independent of the type of superposition used in the key register. We consider this type of structure to be the natural setup for search. We confirm the validity of our new approach through experimental results obtained on real quantum hardware, the Honeywell System Model H0 trapped-ion quantum computer.
△ Less
Submitted 21 July, 2020;
originally announced July 2020.
-
Canonical Construction of Quantum Oracles
Authors:
Austin Gilliam,
Marco Pistoia,
Constantin Gonciulea
Abstract:
Selecting a set of basis states is a common task in quantum computing, in order to increase and/or evaluate their probabilities. This is similar to designing WHERE clauses in classical database queries. Even though one can find heuristic methods to achieve this, it is desirable to automate the process. A common, but inefficient automation approach is to use oracles with classical evaluation of all…
▽ More
Selecting a set of basis states is a common task in quantum computing, in order to increase and/or evaluate their probabilities. This is similar to designing WHERE clauses in classical database queries. Even though one can find heuristic methods to achieve this, it is desirable to automate the process. A common, but inefficient automation approach is to use oracles with classical evaluation of all the states at circuit design time. In this paper, we present a novel, canonical way to produce a quantum oracle from an algebraic expression (in particular, an Ising model), that maps a set of selected states to the same value, coupled with a simple oracle that matches that particular value. We also introduce a general form of the Grover iterate that standardizes this type of oracle. We then apply this new methodology to particular cases of Ising Hamiltonians that model the zero-sum subset problem and the computation of Fibonacci numbers. In addition, this paper presents experimental results obtained on real quantum hardware, the new Honeywell computer based on trapped-ion technology with quantum volume 64.
△ Less
Submitted 18 June, 2020;
originally announced June 2020.
-
Optimizing Quantum Search Using a Generalized Version of Grover's Algorithm
Authors:
Austin Gilliam,
Marco Pistoia,
Constantin Gonciulea
Abstract:
Grover's Search algorithm was a breakthrough at the time it was introduced, and its underlying procedure of amplitude amplification has been a building block of many other algorithms and patterns for extracting information encoded in quantum states. In this paper, we introduce an optimization of the inversion-by-the-mean step of the algorithm. This optimization serves two purposes: from a practica…
▽ More
Grover's Search algorithm was a breakthrough at the time it was introduced, and its underlying procedure of amplitude amplification has been a building block of many other algorithms and patterns for extracting information encoded in quantum states. In this paper, we introduce an optimization of the inversion-by-the-mean step of the algorithm. This optimization serves two purposes: from a practical perspective, it can lead to a performance improvement; from a theoretical one, it leads to a novel interpretation of the actual nature of this step. This step is a reflection, which is realized by (a) cancelling the superposition of a general state to revert to the original all-zeros state, (b) flip** the sign of the amplitude of the all-zeros state, and finally (c) reverting back to the superposition state. Rather than canceling the superposition, our approach allows for going forward to another state that makes the reflection easier. We validate our approach on set and array search, and confirm our results experimentally on real quantum hardware.
△ Less
Submitted 26 May, 2020; v1 submitted 13 May, 2020;
originally announced May 2020.
-
Grover Adaptive Search for Constrained Polynomial Binary Optimization
Authors:
Austin Gilliam,
Stefan Woerner,
Constantin Gonciulea
Abstract:
In this paper we discuss Grover Adaptive Search (GAS) for Constrained Polynomial Binary Optimization (CPBO) problems, and in particular, Quadratic Unconstrained Binary Optimization (QUBO) problems, as a special case. GAS can provide a quadratic speed-up for combinatorial optimization problems compared to brute force search. However, this requires the development of efficient oracles to represent p…
▽ More
In this paper we discuss Grover Adaptive Search (GAS) for Constrained Polynomial Binary Optimization (CPBO) problems, and in particular, Quadratic Unconstrained Binary Optimization (QUBO) problems, as a special case. GAS can provide a quadratic speed-up for combinatorial optimization problems compared to brute force search. However, this requires the development of efficient oracles to represent problems and flag states that satisfy certain search criteria. In general, this can be achieved using quantum arithmetic, however, this is expensive in terms of Toffoli gates as well as required ancilla qubits, which can be prohibitive in the near-term. Within this work, we develop a way to construct efficient oracles to solve CPBO problems using GAS algorithms. We demonstrate this approach and the potential speed-up for the portfolio optimization problem, i.e. a QUBO, using simulation and experimental results obtained on real quantum hardware. However, our approach applies to higher-degree polynomial objective functions as well as constrained optimization problems.
△ Less
Submitted 6 April, 2021; v1 submitted 9 December, 2019;
originally announced December 2019.
-
Foundational Patterns for Efficient Quantum Computing
Authors:
Austin Gilliam,
Charlene Venci,
Sreraman Muralidharan,
Vitaliy Dorum,
Eric May,
Rajesh Narasimhan,
Constantin Gonciulea
Abstract:
We present a number of quantum computing patterns that build on top of fundamental algorithms, that can be applied to solving concrete, NP-hard problems. In particular, we introduce the concept of a quantum dictionary as a summation of multiple patterns and algorithms, and show how it can be applied in the context of Quadratic Unconstrained Binary Optimization (QUBO) problems. We start by presenti…
▽ More
We present a number of quantum computing patterns that build on top of fundamental algorithms, that can be applied to solving concrete, NP-hard problems. In particular, we introduce the concept of a quantum dictionary as a summation of multiple patterns and algorithms, and show how it can be applied in the context of Quadratic Unconstrained Binary Optimization (QUBO) problems. We start by presenting a visual approach to quantum computing, which avoids a heavy-reliance on quantum mechanics, linear algebra, or complex mathematical notation, and favors geometrical intuition and computing paradigms. We also provide insights on the fundamental quantum computing algorithms (Fourier Transforms, Phase Estimation, Grover, Quantum Counting, and Amplitude Estimation).
△ Less
Submitted 25 January, 2021; v1 submitted 23 July, 2019;
originally announced July 2019.
-
Learning Dense Stereo Matching for Digital Surface Models from Satellite Imagery
Authors:
Wayne Treible,
Scott Sorensen,
Andrew D. Gilliam,
Chandra Kambhamettu,
Joseph L. Mundy
Abstract:
Digital Surface Model generation from satellite imagery is a difficult task that has been largely overlooked by the deep learning community. Stereo reconstruction techniques developed for terrestrial systems including self driving cars do not translate well to satellite imagery where image pairs vary considerably. In this work we present neural network tailored for Digital Surface Model generation…
▽ More
Digital Surface Model generation from satellite imagery is a difficult task that has been largely overlooked by the deep learning community. Stereo reconstruction techniques developed for terrestrial systems including self driving cars do not translate well to satellite imagery where image pairs vary considerably. In this work we present neural network tailored for Digital Surface Model generation, a ground truthing and training scheme which maximizes available hardware, and we present a comparison to existing methods. The resulting models are smooth, preserve boundaries, and enable further processing. This represents one of the first attempts at leveraging deep learning in this domain.
△ Less
Submitted 11 December, 2018; v1 submitted 8 November, 2018;
originally announced November 2018.
