Matthew Bowers
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Research Interests
Matthew Bowers has a total of 13 co-authors. Here are the top co-authors by number of shared publications:
- Armando Solar-Lezama
Massachusetts Institute of Technology
2 shared publications - Joshua B. Tenenbaum
Massachusetts Institute of Technology
2 shared publications A. Cacciuto
Department of Chemistry, Columbia University, 3000 Broadway, New York, New York 10027, USA
1 shared publication- Alexander K. Lew
Yale University, New Haven, USA
1 shared publication - Angelo Cacciuto
Columbia University
1 shared publication - Catherine Wong
Massachusetts Institute of Technology, USA
1 shared publication
Latest Publications
- Open Access
- 10 June 2025
- Published by Association for Computing Machinery (ACM) in Proceedings of the ACM on Programming Languages
- Vol. 9 (PLDI) , 1863-1887
Stochastic Lazy Knowledge Compilation for Inference in Discrete Probabilistic Programs
We present new techniques for exact and approximate inference in discrete probabilistic programs, based on two new ways of exploiting lazy evaluation. First, we show how knowledge compilation, a state-of-the art technique for exact inference in discrete probabilistic programs, can be made lazy, enabling asymptotic speed-ups. Second, we show how a probabilistic program’s lazy semantics naturally give rise to a division of its random choices into subproblems, which can be solved in sequence by sequential Monte Carlo with locally-optimal proposals automatically computed via lazy knowledge compilation. We implement our approach in a new tool, Pluck, and evaluate its performance against state-of-the-art approaches to inference in discrete probabilistic languages. We find that on a suite of inference benchmarks, lazy knowledge compilation can be faster than state-of-the-art approaches, sometimes by orders of magnitude.
- 10 June 2025
- Published by Association for Computing Machinery (ACM) in Proceedings of the ACM on Programming Languages
- Vol. 9 (PLDI) , 1863-1887
Open AccessStochastic Lazy Knowledge Compilation for Inference in Discrete Probabilistic Programs
We present new techniques for exact and approximate inference in discrete probabilistic programs, based on two new ways of exploiting lazy evaluation. First, we show how knowledge compilation, a state-of-the art technique for exact inference in discrete probabilistic programs, can be made lazy, enabling asymptotic speed-ups. Second, we show how a probabilistic program’s lazy semantics naturally give rise to a division of its random choices into subproblems, which can be solved in sequence by sequential Monte Carlo with locally-optimal proposals automatically computed via lazy knowledge compilation. We implement our approach in a new tool, Pluck, and evaluate its performance against state-of-the-art approaches to inference in discrete probabilistic languages. We find that on a suite of inference benchmarks, lazy knowledge compilation can be faster than state-of-the-art approaches, sometimes by orders of magnitude.
- Open Access
- 9 January 2023
- Published by Association for Computing Machinery (ACM) in Proceedings of the ACM on Programming Languages
- Vol. 7 (POPL) , 1182-1213
Top-Down Synthesis for Library Learning
This paper introduces corpus-guided top-down synthesis as a mechanism for synthesizing library functions that capture common functionality from a corpus of programs in a domain specific language (DSL). The algorithm builds abstractions directly from initial DSL primitives, using syntactic pattern matching of intermediate abstractions to intelligently prune the search space and guide the algorithm towards abstractions that maximally capture shared structures in the corpus. We present an implementation of the approach in a tool called Stitch and evaluate it against the state-of-the-art deductive library learning algorithm from DreamCoder. Our evaluation shows that Stitch is 3-4 orders of magnitude faster and uses 2 orders of magnitude less memory while maintaining comparable or better library quality (as measured by compressivity). We also demonstrate Stitch’s scalability on corpora containing hundreds of complex programs that are intractable with prior deductive approaches and show empirically that it is robust to terminating the search procedure early—further allowing it to scale to challenging datasets by means of early stopping.
- 9 January 2023
- Published by Association for Computing Machinery (ACM) in Proceedings of the ACM on Programming Languages
- Vol. 7 (POPL) , 1182-1213
Open AccessTop-Down Synthesis for Library Learning
This paper introduces corpus-guided top-down synthesis as a mechanism for synthesizing library functions that capture common functionality from a corpus of programs in a domain specific language (DSL). The algorithm builds abstractions directly from initial DSL primitives, using syntactic pattern matching of intermediate abstractions to intelligently prune the search space and guide the algorithm towards abstractions that maximally capture shared structures in the corpus. We present an implementation of the approach in a tool called Stitch and evaluate it against the state-of-the-art deductive library learning algorithm from DreamCoder. Our evaluation shows that Stitch is 3-4 orders of magnitude faster and uses 2 orders of magnitude less memory while maintaining comparable or better library quality (as measured by compressivity). We also demonstrate Stitch’s scalability on corpora containing hundreds of complex programs that are intractable with prior deductive approaches and show empirically that it is robust to terminating the search procedure early—further allowing it to scale to challenging datasets by means of early stopping.
- 24 August 2020
- Published by AIP Publishing in The Journal of Chemical Physics
- Vol. 153 (8) , 084901
Universal reshaping of arrested colloidal gels via active doping
Colloids that interact via a short-range attraction serve as the primary building blocks for a broad range of self-assembled materials. However, one of the well-known drawbacks to this strategy is that these building blocks rapidly and readily condense into a metastable colloidal gel. Using computer simulations, we illustrate how the addition of a small fraction of purely repulsive self-propelled colloids, a technique referred to as active doping, can prevent the formation of this metastable gel state and drive the system toward its thermodynamically favored crystalline target structure. The simplicity and robust nature of this strategy offers a systematic and generic pathway to improving the self-assembly of a large number of complex colloidal structures. We discuss in detail the process by which this feat is accomplished and provide quantitative metrics for exploiting it to modulate the self-assembly. We provide evidence for the generic nature of this approach by demonstrating that it remains robust under a number of different anisotropic short-ranged pair interactions in both two and three dimensions. In addition, we report on a novel microphase in mixtures of passive and active colloids. For a broad range of self-propelling velocities, it is possible to stabilize a suspension of fairly monodisperse finite-size crystallites. Surprisingly, this microphase is also insensitive to the underlying pair interaction between building blocks. The active stabilization of these moderately sized monodisperse clusters is quite remarkable and should be of great utility in the design of hierarchical self-assembly strategies. This work further bolsters the notion that active forces can play a pivotal role in directing colloidal self-assembly.
- 24 August 2020
- Published by AIP Publishing in The Journal of Chemical Physics
- Vol. 153 (8) , 084901
Universal reshaping of arrested colloidal gels via active doping
Colloids that interact via a short-range attraction serve as the primary building blocks for a broad range of self-assembled materials. However, one of the well-known drawbacks to this strategy is that these building blocks rapidly and readily condense into a metastable colloidal gel. Using computer simulations, we illustrate how the addition of a small fraction of purely repulsive self-propelled colloids, a technique referred to as active doping, can prevent the formation of this metastable gel state and drive the system toward its thermodynamically favored crystalline target structure. The simplicity and robust nature of this strategy offers a systematic and generic pathway to improving the self-assembly of a large number of complex colloidal structures. We discuss in detail the process by which this feat is accomplished and provide quantitative metrics for exploiting it to modulate the self-assembly. We provide evidence for the generic nature of this approach by demonstrating that it remains robust under a number of different anisotropic short-ranged pair interactions in both two and three dimensions. In addition, we report on a novel microphase in mixtures of passive and active colloids. For a broad range of self-propelling velocities, it is possible to stabilize a suspension of fairly monodisperse finite-size crystallites. Surprisingly, this microphase is also insensitive to the underlying pair interaction between building blocks. The active stabilization of these moderately sized monodisperse clusters is quite remarkable and should be of great utility in the design of hierarchical self-assembly strategies. This work further bolsters the notion that active forces can play a pivotal role in directing colloidal self-assembly.
- Open Access
- 4 April 2019
- Published by AIP Publishing in The Journal of Chemical Physics
- Vol. 150 (13) , 134505
Active sculpting of colloidal crystals
The development of colloidal particles that can become self-propelled when exposed to a source of light of a given frequency represents one of the most exciting new developments in the field of active matter. In this paper, we perform numerical simulations to explore several properties of crystalline colloidal aggregates in the presence of such external light fields. We show how permanent gaps of tunable size can be carved into these crystals as a result of a solid-gas transition that occurs above a threshold light intensity. We compare this phenomenon to that occurring in a parent system obtained by replacing self-propulsion with an effective temperature and discuss the main differences between the two setups. Finally, we show how moving the light field across a solid-fluid boundary allows us to transfer or transport the fluid component into or across the solid one. We discuss the dynamics of this process and set limits associated with its performance.
- 4 April 2019
- Published by AIP Publishing in The Journal of Chemical Physics
- Vol. 150 (13) , 134505
Open AccessActive sculpting of colloidal crystals
The development of colloidal particles that can become self-propelled when exposed to a source of light of a given frequency represents one of the most exciting new developments in the field of active matter. In this paper, we perform numerical simulations to explore several properties of crystalline colloidal aggregates in the presence of such external light fields. We show how permanent gaps of tunable size can be carved into these crystals as a result of a solid-gas transition that occurs above a threshold light intensity. We compare this phenomenon to that occurring in a parent system obtained by replacing self-propulsion with an effective temperature and discuss the main differences between the two setups. Finally, we show how moving the light field across a solid-fluid boundary allows us to transfer or transport the fluid component into or across the solid one. We discuss the dynamics of this process and set limits associated with its performance.