Showing 1–44 of 44 results for author: Lew, A

Probabilistic Programming with Programmable Variational Inference

Authors: McCoy R. Becker, Alexander K. Lew, Xiaoyan Wang, Matin Ghavami, Mathieu Huot, Martin C. Rinard, Vikash K. Mansinghka

Abstract: Compared to the wide array of advanced Monte Carlo methods supported by modern probabilistic programming languages (PPLs), PPL support for variational inference (VI) is less developed: users are typically limited to a predefined selection of variational objectives and gradient estimators, which are implemented monolithically (and without formal correctness arguments) in PPL backends. In this paper… ▽ More Compared to the wide array of advanced Monte Carlo methods supported by modern probabilistic programming languages (PPLs), PPL support for variational inference (VI) is less developed: users are typically limited to a predefined selection of variational objectives and gradient estimators, which are implemented monolithically (and without formal correctness arguments) in PPL backends. In this paper, we propose a more modular approach to supporting variational inference in PPLs, based on compositional program transformation. In our approach, variational objectives are expressed as programs, that may employ first-class constructs for computing densities of and expected values under user-defined models and variational families. We then transform these programs systematically into unbiased gradient estimators for optimizing the objectives they define. Our design enables modular reasoning about many interacting concerns, including automatic differentiation, density accumulation, tracing, and the application of unbiased gradient estimation strategies. Additionally, relative to existing support for VI in PPLs, our design increases expressiveness along three axes: (1) it supports an open-ended set of user-defined variational objectives, rather than a fixed menu of options; (2) it supports a combinatorial space of gradient estimation strategies, many not automated by today's PPLs; and (3) it supports a broader class of models and variational families, because it supports constructs for approximate marginalization and normalization (previously introduced only for Monte Carlo inference). We implement our approach in an extension to the Gen probabilistic programming system (genjax.vi, implemented in JAX), and evaluate on several deep generative modeling tasks, showing minimal performance overhead vs. hand-coded implementations and performance competitive with well-established open-source PPLs. △ Less

Submitted 22 June, 2024; originally announced June 2024.

Journal ref: PLDI 2024

GenSQL: A Probabilistic Programming System for Querying Generative Models of Database Tables

Authors: Mathieu Huot, Matin Ghavami, Alexander K. Lew, Ulrich Schaechtle, Cameron E. Freer, Zane Shelby, Martin C. Rinard, Feras A. Saad, Vikash K. Mansinghka

Abstract: This article presents GenSQL, a probabilistic programming system for querying probabilistic generative models of database tables. By augmenting SQL with only a few key primitives for querying probabilistic models, GenSQL enables complex Bayesian inference workflows to be concisely implemented. GenSQL's query planner rests on a unified programmatic interface for interacting with probabilistic model… ▽ More This article presents GenSQL, a probabilistic programming system for querying probabilistic generative models of database tables. By augmenting SQL with only a few key primitives for querying probabilistic models, GenSQL enables complex Bayesian inference workflows to be concisely implemented. GenSQL's query planner rests on a unified programmatic interface for interacting with probabilistic models of tabular data, which makes it possible to use models written in a variety of probabilistic programming languages that are tailored to specific workflows. Probabilistic models may be automatically learned via probabilistic program synthesis, hand-designed, or a combination of both. GenSQL is formalized using a novel type system and denotational semantics, which together enable us to establish proofs that precisely characterize its soundness guarantees. We evaluate our system on two case real-world studies -- an anomaly detection in clinical trials and conditional synthetic data generation for a virtual wet lab -- and show that GenSQL more accurately captures the complexity of the data as compared to common baselines. We also show that the declarative syntax in GenSQL is more concise and less error-prone as compared to several alternatives. Finally, GenSQL delivers a 1.7-6.8x speedup compared to its closest competitor on a representative benchmark set and runs in comparable time to hand-written code, in part due to its reusable optimizations and code specialization. △ Less

Submitted 21 June, 2024; originally announced June 2024.

Comments: 54 pages, 30 figures, 1 table, published at PLDI 2024

Simons Observatory: Pre-deployment Performance of a Large Aperture Telescope Optics Tube in the 90 and 150 GHz Spectral Bands

Authors: Carlos E. Sierra, Kathleen Harrington, Shreya Sutariya, Thomas Alford, Anna M. Kofman, Grace E. Chesmore, Jason E. Austermann, Andrew Bazarko, James A. Beall, Tanay Bhandarkar, Mark J. Devlin, Simon R. Dicker, Peter N. Dow, Shannon M. Duff, Daniel Dutcher, Nicholas Galitzki, Joseph E. Golec, John C. Groh, Jon E. Gudmundsson, Saianeesh K. Haridas, Erin Healy, Johannes Hubmayr, Jeffrey Iuliano, Bradley R. Johnson, Claire S. Lessler , et al. (20 additional authors not shown)

Abstract: The Simons Observatory will map the temperature and polarization over half of the sky, at millimeter wavelengths in six spectral bands from the Atacama Desert in Chile. These data will provide new insights into the genesis, content, and history of our Universe; the astrophysics of galaxies and galaxy clusters; objects in our solar system; and time-varying astrophysical phenomena. This ambitious ne… ▽ More The Simons Observatory will map the temperature and polarization over half of the sky, at millimeter wavelengths in six spectral bands from the Atacama Desert in Chile. These data will provide new insights into the genesis, content, and history of our Universe; the astrophysics of galaxies and galaxy clusters; objects in our solar system; and time-varying astrophysical phenomena. This ambitious new instrument suite, initially comprising three 0.5 m small-aperture telescopes and one 6 m large aperture telescope, is designed using a common combination of new technologies and new implementations to realize an observatory significantly more capable than the previous generation. In this paper, we present the pre-deployment performance of the first mid-frequency "optics tube" which will be fielded on the large aperture telescope with sensitivity to the 90 and 150 GHz spectral bands. This optics tube contains lenses, filters, detectors, and readout components, all of which operate at cryogenic temperatures. It is one of seven that form the core of the large aperture telescope receiver in its initial deployment. We describe this optics tube, including details of comprehensive testing methods, new techniques for beam and passband characterization, and its measured performance. The performance metrics include beams, optical efficiency, passbands, and forecasts for the on-sky performance of the system. We forecast a sensitivity that exceeds the requirements of the large aperture telescope with greater than 30% margin in each spectral band, and predict that the instrument will realize diffraction-limited performance and the expected detector passbands. △ Less

Submitted 10 May, 2024; originally announced May 2024.

The Simons Observatory: Design, integration, and testing of the small aperture telescopes

Authors: Nicholas Galitzki, Tran Tsan, Jake Spisak, Michael Randall, Max Silva-Feaver, Joseph Seibert, Jacob Lashner, Shunsuke Adachi, Sean M. Adkins, Thomas Alford, Kam Arnold, Peter C. Ashton, Jason E. Austermann, Carlo Baccigalupi, Andrew Bazarko, James A. Beall, Sanah Bhimani, Bryce Bixler, Gabriele Coppi, Lance Corbett, Kevin D. Crowley, Kevin T. Crowley, Samuel Day-Weiss, Simon Dicker, Peter N. Dow , et al. (55 additional authors not shown)

Abstract: The Simons Observatory (SO) is a cosmic microwave background (CMB) survey experiment that includes small-aperture telescopes (SATs) observing from an altitude of 5,200 m in the Atacama Desert in Chile. The SO SATs will cover six spectral bands between 27 and 280 GHz to search for primordial B-modes to a sensitivity of $σ(r)=0.002$, with quantified systematic errors well below this value. Each SAT… ▽ More The Simons Observatory (SO) is a cosmic microwave background (CMB) survey experiment that includes small-aperture telescopes (SATs) observing from an altitude of 5,200 m in the Atacama Desert in Chile. The SO SATs will cover six spectral bands between 27 and 280 GHz to search for primordial B-modes to a sensitivity of $σ(r)=0.002$, with quantified systematic errors well below this value. Each SAT is a self-contained cryogenic telescope with a 35$^\circ$ field of view, 42 cm diameter optical aperture, 40 K half-wave plate, 1 K refractive optics, and $<0.1$ K focal plane that holds $>12,000$ TES detectors. We describe the nominal design of the SATs and present details about the integration and testing for one operating at 93 and 145 GHz. △ Less

Submitted 10 May, 2024; v1 submitted 9 May, 2024; originally announced May 2024.

High Absorptivity Nanotextured Powders for Additive Manufacturing

Authors: Ottman A. Tertuliano, Philip J. DePond, Andrew C. Lee, Jiho Hong, David Doan, Luc Capaldi, Mark Brongersma, X. Wendy Gu, Manyalibo J. Matthews, Wei Cai, Adrian J. Lew

Abstract: The widespread application of metal additive manufacturing (AM) is limited by the ability to control the complex interactions between the energy source and the feedstock material. Here we develop a generalizable process to introduce nanoscale grooves to the surface of metal powders which increases the powder absorptivity by up to 70% during laser powder bed fusion. Absorptivity enhancements in cop… ▽ More The widespread application of metal additive manufacturing (AM) is limited by the ability to control the complex interactions between the energy source and the feedstock material. Here we develop a generalizable process to introduce nanoscale grooves to the surface of metal powders which increases the powder absorptivity by up to 70% during laser powder bed fusion. Absorptivity enhancements in copper, copper-silver, and tungsten enables energy efficient manufacturing, with printing of pure copper at relative densities up to 92% using laser energy densities as low as 82 J/mm^3. Simulations show the enhanced powder absorptivity results from plasmon-enabled light concentration in nanoscale grooves combined with multiple scattering events. The approach taken here demonstrates a general method to enhance the absorptivity and printability of reflective and refractory metal powders by changing the surface morphology of the feedstock without altering its composition. △ Less

Submitted 8 December, 2023; originally announced December 2023.

Rigid partitions: from high connectivity to random graphs

Authors: Michael Krivelevich, Alan Lew, Peleg Michaeli

Abstract: A graph is called $d$-rigid if there exists a generic embedding of its vertex set into $\mathbb{R}^d$ such that every continuous motion of the vertices that preserves the lengths of all edges actually preserves the distances between all pairs of vertices. The rigidity of a graph is the maximal $d$ such that the graph is $d$-rigid. We present new sufficient conditions for the $d$-rigidity of a grap… ▽ More A graph is called $d$-rigid if there exists a generic embedding of its vertex set into $\mathbb{R}^d$ such that every continuous motion of the vertices that preserves the lengths of all edges actually preserves the distances between all pairs of vertices. The rigidity of a graph is the maximal $d$ such that the graph is $d$-rigid. We present new sufficient conditions for the $d$-rigidity of a graph in terms of the existence of ``rigid partitions'' -- partitions of the graph that satisfy certain connectivity properties. This extends previous results by Crapo, Lindemann, and Lew, Nevo, Peled and Raz. As an application, we present new results on the rigidity of highly-connected graphs, random graphs, random bipartite graphs, pseudorandom graphs, and dense graphs. In particular, we prove that random $C d\log d$-regular graphs are typically $d$-rigid, demonstrate the existence of a giant $d$-rigid component in sparse random binomial graphs, and show that the rigidity of relatively sparse random binomial bipartite graphs is roughly the same as that of the complete bipartite graph, which we consider an interesting phenomenon. Furthermore, we show that a graph admitting $\binom{d+1}{2}$ disjoint connected dominating sets is $d$-rigid. This implies a weak version of the Lovász--Yemini conjecture on the rigidity of highly-connected graphs. We also present an alternative short proof for a recent result by Lew, Nevo, Peled, and Raz, which asserts that the hitting time for $d$-rigidity in the random graph process typically coincides with the hitting time for minimum degree $d$. △ Less

Submitted 12 December, 2023; v1 submitted 24 November, 2023; originally announced November 2023.

Comments: 30 pages. In this updated version, we have added a theorem concerning the rigidity of dense graphs and incorporated references to Villányi's recent resolution of the Lovász-Yemini conjecture

MSC Class: 05C10; 52C25; 05C40; 05C80; 05C50

Asymptotic behavior of Laplacian eigenvalues of subspace inclusion graphs

Authors: Alan Lew

Abstract: Let $\text{Fl}_{n,q}$ be the simplicial complex whose vertices are the non-trivial subspaces of $\mathbb{F}_q^n$ and whose simplices correspond to families of subspaces forming a flag. Let $Δ^{+}_k(\text{Fl}_{n,q})$ be the $k$-dimensional weighted upper Laplacian on $ \text{Fl}_{n,q}$. The spectrum of $Δ^{+}_k(\text{Fl}_{n,q})$ was first studied by Garland, who obtained a lower bound on its non-ze… ▽ More Let $\text{Fl}_{n,q}$ be the simplicial complex whose vertices are the non-trivial subspaces of $\mathbb{F}_q^n$ and whose simplices correspond to families of subspaces forming a flag. Let $Δ^{+}_k(\text{Fl}_{n,q})$ be the $k$-dimensional weighted upper Laplacian on $ \text{Fl}_{n,q}$. The spectrum of $Δ^{+}_k(\text{Fl}_{n,q})$ was first studied by Garland, who obtained a lower bound on its non-zero eigenvalues. Here, we focus on the $k=0$ case. We determine the asymptotic behavior of the eigenvalues of $Δ_{0}^{+}(\text{Fl}_{n,q})$ as $q$ tends to infinity. In particular, we show that for large enough $q$, $Δ_{0}^{+}(\text{Fl}_{n,q})$ has exactly $\left\lfloor n^2/4\right\rfloor+2$ distinct eigenvalues, and that every eigenvalue $λ\neq 0,n-1$ of $Δ_{0}^{+}(\text{Fl}_{n,q})$ tends to $n-2$ as $q$ goes to infinity. This solves the $0$-dimensional case of a conjecture of Papikian. △ Less

Submitted 16 August, 2023; originally announced August 2023.

Laplacian eigenvalues of independence complexes via additive compound matrices

Authors: Alan Lew

Abstract: The independence complex of a graph $G=(V,E)$ is the simplicial complex $I(G)$ on vertex set $V$ whose simplices are the independent sets in $G$. We present new lower bounds on the eigenvalues of the $k$-dimensional Laplacian $L_k(I(G))$ in terms of the eigenvalues of the graph Laplacian $L(G)$. As a consequence, we show that for all $k\geq 0$, the dimension of the $k$-th reduced homology group (w… ▽ More The independence complex of a graph $G=(V,E)$ is the simplicial complex $I(G)$ on vertex set $V$ whose simplices are the independent sets in $G$. We present new lower bounds on the eigenvalues of the $k$-dimensional Laplacian $L_k(I(G))$ in terms of the eigenvalues of the graph Laplacian $L(G)$. As a consequence, we show that for all $k\geq 0$, the dimension of the $k$-th reduced homology group (with real coefficients) of $I(G)$ is at most \[ \left| \left\{ 1\leq i_1<\cdots<i_{k+1}\leq |V| : \, λ_{i_1}+λ_{i_2}+\cdots+λ_{i_{k+1}} \geq |V|\right\}\right|,\] where $λ_1\geqλ_2\geq \cdots\geq λ_{|V|}=0$ are the eigenvalues of $L(G)$. In particular, if $k$ is the minimal number such that the sum of the $k$ largest eigenvalues of $L(G)$ is at least $|V|$, then $\tilde{H}_i(I(G);\mathbb{R})=0$ for all $i\leq k-2$. This extends previous results by Aharoni, Berger and Meshulam. Our proof relies on a relation between the $k$-dimensional Laplacian $L_k(I(G))$ and the $(k+1)$-th additive compound matrix of $L_0(I(G))$, which is an $\binom{n}{k+1}\times\binom{n}{k+1}$ matrix whose eigenvalues are all the possible sums of $k+1$ eigenvalues of the $0$-dimensional Laplacian. Our results apply also in the more general setting of vertex-weighted Laplacian matrices. △ Less

Submitted 26 July, 2023; originally announced July 2023.

A finite element method to compute the dam** rate and frequency of oscillating fluids inside microfluidic nozzles

Authors: Søren Taverniers, Svyatoslav Korneev, Christoforos Somarakis, Morad Behandish, Adrian J. Lew

Abstract: The computation of dam** rates of an oscillating fluid with a free surface in which viscosity is small and surface tension high is numerically challenging. A typical application requiring such computation is drop-on-demand (DoD) microfluidic devices that eject liquid metal droplets, where accurate knowledge of dam** rates for the least-damped oscillation modes following droplet ejection is par… ▽ More The computation of dam** rates of an oscillating fluid with a free surface in which viscosity is small and surface tension high is numerically challenging. A typical application requiring such computation is drop-on-demand (DoD) microfluidic devices that eject liquid metal droplets, where accurate knowledge of dam** rates for the least-damped oscillation modes following droplet ejection is paramount for assessing jetting stability at higher jetting frequencies. Computational fluid dynamics (CFD) simulations often struggle to accurately predict meniscus dam** unless a very fine discretization is adopted, so calculations are computationally expensive. The faster alternative we adopt is to compute dam** rates directly from the eigenvalues of the linearized problem. The surface tension term in Stokes or sloshing problems requires approximation of meniscus displacements, which introduces additional complexity in their numerical solution. We consider the combined effects of viscosity and surface tension, approximate the meniscus displacements, and construct a finite element method to compute the fluid's oscillation modes. We prove that the method is free of spurious modes with zero or positive dam** rates, and we implement it with Taylor-Hood elements for velocity and pressure, and with continuous piecewise quadratic elements for meniscus displacement. We verify the numerical convergence of the method by reproducing the solution to an analytical benchmark problem and two more complex examples with axisymmetric geometry. We obtain the spatial shape and temporal evolution (angular frequency and dam** rate) of the set of least-damped oscillation modes in minutes, compared to days for a CFD simulation. The method's ability to quickly generate accurate estimates of fluid oscillation dam** rates makes it suitable for integration into design loops for prototy** microfluidic nozzles. △ Less

Submitted 10 June, 2024; v1 submitted 30 June, 2023; originally announced July 2023.

From Word Models to World Models: Translating from Natural Language to the Probabilistic Language of Thought

Authors: Lionel Wong, Gabriel Grand, Alexander K. Lew, Noah D. Goodman, Vikash K. Mansinghka, Jacob Andreas, Joshua B. Tenenbaum

Abstract: How does language inform our downstream thinking? In particular, how do humans make meaning from language--and how can we leverage a theory of linguistic meaning to build machines that think in more human-like ways? In this paper, we propose rational meaning construction, a computational framework for language-informed thinking that combines neural language models with probabilistic models for rat… ▽ More How does language inform our downstream thinking? In particular, how do humans make meaning from language--and how can we leverage a theory of linguistic meaning to build machines that think in more human-like ways? In this paper, we propose rational meaning construction, a computational framework for language-informed thinking that combines neural language models with probabilistic models for rational inference. We frame linguistic meaning as a context-sensitive map** from natural language into a probabilistic language of thought (PLoT)--a general-purpose symbolic substrate for generative world modeling. Our architecture integrates two computational tools that have not previously come together: we model thinking with probabilistic programs, an expressive representation for commonsense reasoning; and we model meaning construction with large language models (LLMs), which support broad-coverage translation from natural language utterances to code expressions in a probabilistic programming language. We illustrate our framework through examples covering four core domains from cognitive science: probabilistic reasoning, logical and relational reasoning, visual and physical reasoning, and social reasoning. In each, we show that LLMs can generate context-sensitive translations that capture pragmatically-appropriate linguistic meanings, while Bayesian inference with the generated programs supports coherent and robust commonsense reasoning. We extend our framework to integrate cognitively-motivated symbolic modules (physics simulators, graphics engines, and planning algorithms) to provide a unified commonsense thinking interface from language. Finally, we explore how language can drive the construction of world models themselves. We hope this work will provide a roadmap towards cognitive models and AI systems that synthesize the insights of both modern and classical computational perspectives. △ Less

Submitted 23 June, 2023; v1 submitted 22 June, 2023; originally announced June 2023.

Differentiating Metropolis-Hastings to Optimize Intractable Densities

Authors: Gaurav Arya, Ruben Seyer, Frank Schäfer, Kartik Chandra, Alexander K. Lew, Mathieu Huot, Vikash K. Mansinghka, Jonathan Ragan-Kelley, Christopher Rackauckas, Moritz Schauer

Abstract: We develop an algorithm for automatic differentiation of Metropolis-Hastings samplers, allowing us to differentiate through probabilistic inference, even if the model has discrete components within it. Our approach fuses recent advances in stochastic automatic differentiation with traditional Markov chain coupling schemes, providing an unbiased and low-variance gradient estimator. This allows us t… ▽ More We develop an algorithm for automatic differentiation of Metropolis-Hastings samplers, allowing us to differentiate through probabilistic inference, even if the model has discrete components within it. Our approach fuses recent advances in stochastic automatic differentiation with traditional Markov chain coupling schemes, providing an unbiased and low-variance gradient estimator. This allows us to apply gradient-based optimization to objectives expressed as expectations over intractable target densities. We demonstrate our approach by finding an ambiguous observation in a Gaussian mixture model and by maximizing the specific heat in an Ising model. △ Less

Submitted 30 June, 2023; v1 submitted 13 June, 2023; originally announced June 2023.

Comments: 6 pages, 6 figures; accepted at Differentiable Almost Everything Workshop of ICML 2023

Sequential Monte Carlo Steering of Large Language Models using Probabilistic Programs

Authors: Alexander K. Lew, Tan Zhi-Xuan, Gabriel Grand, Vikash K. Mansinghka

Abstract: Even after fine-tuning and reinforcement learning, large language models (LLMs) can be difficult, if not impossible, to control reliably with prompts alone. We propose a new inference-time approach to enforcing syntactic and semantic constraints on the outputs of LLMs, called sequential Monte Carlo (SMC) steering. The key idea is to specify language generation tasks as posterior inference problems… ▽ More Even after fine-tuning and reinforcement learning, large language models (LLMs) can be difficult, if not impossible, to control reliably with prompts alone. We propose a new inference-time approach to enforcing syntactic and semantic constraints on the outputs of LLMs, called sequential Monte Carlo (SMC) steering. The key idea is to specify language generation tasks as posterior inference problems in a class of discrete probabilistic sequence models, and replace standard decoding with sequential Monte Carlo inference. For a computational cost similar to that of beam search, SMC can steer LLMs to solve diverse tasks, including infilling, generation under syntactic constraints, and prompt intersection. To facilitate experimentation with SMC steering, we present a probabilistic programming library, LLaMPPL (https://github.com/probcomp/hfppl), for concisely specifying new generation tasks as language model probabilistic programs, and automating steering of LLaMA-family Transformers. △ Less

Submitted 26 November, 2023; v1 submitted 5 June, 2023; originally announced June 2023.

Comments: Minor typo fixes

Extensions of the Colorful Helly Theorem for $d$-collapsible and $d$-Leray complexes

Authors: Minki Kim, Alan Lew

Abstract: We present extensions of the Colorful Helly Theorem for $d$-collapsible and $d$-Leray complexes, providing a common generalization to the matroidal versions of the theorem due to Kalai and Meshulam, the ``very colorful" Helly theorem introduced by Arocha, Bárány, Bracho, Fabila and Montejano, and the ``semi-intersecting" colorful Helly theorem proved by Montejano and Karasev. As an application,… ▽ More We present extensions of the Colorful Helly Theorem for $d$-collapsible and $d$-Leray complexes, providing a common generalization to the matroidal versions of the theorem due to Kalai and Meshulam, the ``very colorful" Helly theorem introduced by Arocha, Bárány, Bracho, Fabila and Montejano, and the ``semi-intersecting" colorful Helly theorem proved by Montejano and Karasev. As an application, we obtain the following extension of Tverberg's Theorem: Let $A$ be a finite set of points in $\mathbb{R}^d$ with $|A|>(r-1)(d+1)$. Then, there exist a partition $A_1,\ldots,A_r$ of $A$ and a subset $B\subset A$ of size $(r-1)(d+1)$, such that $\cap_{i=1}^r \text{conv}( (B\cup\{p\})\cap A_i)\neq\emptyset$ for all $p\in A\setminus B$. That is, we obtain a partition of $A$ into $r$ parts that remains a Tverberg partition even after removing all but one arbitrary point from $A\setminus B$. △ Less

Submitted 21 May, 2023; originally announced May 2023.

Comments: 20 pages, 1 figure

Garland's method for token graphs

Authors: Alan Lew

Abstract: The $k$-th token graph of a graph $G=(V,E)$ is the graph $F_k(G)$ whose vertices are the $k$-subsets of $V$ and whose edges are all pairs of $k$-subsets $A,B$ such that the symmetric difference of $A$ and $B$ forms an edge in $G$. Let $L(G)$ be the Laplacian matrix of $G$, and $L_k(G)$ be the Laplacian matrix of $F_k(G)$. It was shown by Dalfó et al. that for any graph $G$ on $n$ vertices and any… ▽ More The $k$-th token graph of a graph $G=(V,E)$ is the graph $F_k(G)$ whose vertices are the $k$-subsets of $V$ and whose edges are all pairs of $k$-subsets $A,B$ such that the symmetric difference of $A$ and $B$ forms an edge in $G$. Let $L(G)$ be the Laplacian matrix of $G$, and $L_k(G)$ be the Laplacian matrix of $F_k(G)$. It was shown by Dalfó et al. that for any graph $G$ on $n$ vertices and any $0\leq \ell \leq k \leq \left\lfloor n/2\right\rfloor$, the spectrum of $L_{\ell}(G)$ is contained in that of $L_k(G)$. Here, we continue to study the relation between the spectrum of $L_k(G)$ and that of $L_{k-1}(G)$. In particular, we show that, for $1\leq k\leq \left\lfloor n/2\right\rfloor$, any eigenvalue $λ$ of $L_k(G)$ that is not contained in the spectrum of $L_{k-1}(G)$ satisfies \[ k(λ_2(L(G))-k+1)\leq λ\leq kλ_n(L(G)), \] where $λ_2(L(G))$ is the second smallest eigenvalue of $L(G)$ (a.k.a. the algebraic connectivity of $G$), and $λ_n(L(G))$ is its largest eigenvalue. Our proof relies on an adaptation of Garland's method, originally developed for the study of high-dimensional Laplacians of simplicial complexes. △ Less

Submitted 3 May, 2023; originally announced May 2023.

Rigidity expander graphs

Authors: Alan Lew, Eran Nevo, Yuval Peled, Orit E. Raz

Abstract: Jordán and Tanigawa recently introduced the $d$-dimensional algebraic connectivity $a_d(G)$ of a graph $G$. This is a quantitative measure of the $d$-dimensional rigidity of $G$ which generalizes the well-studied notion of spectral expansion of graphs. We present a new lower bound for $a_d(G)$ defined in terms of the spectral expansion of certain subgraphs of $G$ associated with a partition of its… ▽ More Jordán and Tanigawa recently introduced the $d$-dimensional algebraic connectivity $a_d(G)$ of a graph $G$. This is a quantitative measure of the $d$-dimensional rigidity of $G$ which generalizes the well-studied notion of spectral expansion of graphs. We present a new lower bound for $a_d(G)$ defined in terms of the spectral expansion of certain subgraphs of $G$ associated with a partition of its vertices into $d$ parts. In particular, we obtain a new sufficient condition for the rigidity of a graph $G$. As a first application, we prove the existence of an infinite family of $k$-regular $d$-rigidity-expander graphs for every $d\ge 2$ and $k\ge 2d+1$. Conjecturally, no such family of $2d$-regular graphs exists. Second, we show that $a_d(K_n)\geq \frac{1}{2}\left\lfloor\frac{n}{d}\right\rfloor$, which we conjecture to be essentially tight. In addition, we study the extremal values $a_d(G)$ attained if $G$ is a minimally $d$-rigid graph. △ Less

Submitted 3 April, 2023; originally announced April 2023.

$ω$PAP Spaces: Reasoning Denotationally About Higher-Order, Recursive Probabilistic and Differentiable Programs

Authors: Mathieu Huot, Alexander K. Lew, Vikash K. Mansinghka, Sam Staton

Abstract: We introduce a new setting, the category of $ω$PAP spaces, for reasoning denotationally about expressive differentiable and probabilistic programming languages. Our semantics is general enough to assign meanings to most practical probabilistic and differentiable programs, including those that use general recursion, higher-order functions, discontinuous primitives, and both discrete and continuous… ▽ More We introduce a new setting, the category of $ω$PAP spaces, for reasoning denotationally about expressive differentiable and probabilistic programming languages. Our semantics is general enough to assign meanings to most practical probabilistic and differentiable programs, including those that use general recursion, higher-order functions, discontinuous primitives, and both discrete and continuous sampling. But crucially, it is also specific enough to exclude many pathological denotations, enabling us to establish new results about both deterministic differentiable programs and probabilistic programs. In the deterministic setting, we prove very general correctness theorems for automatic differentiation and its use within gradient descent. In the probabilistic setting, we establish the almost-everywhere differentiability of probabilistic programs' trace density functions, and the existence of convenient base measures for density computation in Monte Carlo inference. In some cases these results were previously known, but required detailed proofs with an operational flavor; by contrast, all our proofs work directly with programs' denotations. △ Less

Submitted 25 May, 2023; v1 submitted 21 February, 2023; originally announced February 2023.

Comments: 11 figures, 10 pages main paper + 13 pages of appendices

Shutter, the Robot Photographer: Leveraging Behavior Trees for Public, In-the-Wild Human-Robot Interactions

Authors: Alexander Lew, Sydney Thompson, Nathan Tsoi, Marynel Vázquez

Abstract: Deploying interactive systems in-the-wild requires adaptability to situations not encountered in lab environments. Our work details our experience about the impact of architecture choice on behavior reusability and reactivity while deploying a public interactive system. In particular, we introduce Shutter, a robot photographer and a platform for public interaction. In designing Shutter's architect… ▽ More Deploying interactive systems in-the-wild requires adaptability to situations not encountered in lab environments. Our work details our experience about the impact of architecture choice on behavior reusability and reactivity while deploying a public interactive system. In particular, we introduce Shutter, a robot photographer and a platform for public interaction. In designing Shutter's architecture, we focused on adaptability for in-the-wild deployment, while develo** a reusable platform to facilitate future research in public human-robot interaction. We find that behavior trees allow reactivity, especially in group settings, and encourage designing reusable behaviors. △ Less

Submitted 31 January, 2023; originally announced February 2023.

Comments: Submitted to the 2022 Human-Robot Interaction in Public Spaces workshop: https://sites.google.com/view/hripublicspaces2022

ADEV: Sound Automatic Differentiation of Expected Values of Probabilistic Programs

Authors: Alexander K. Lew, Mathieu Huot, Sam Staton, Vikash K. Mansinghka

Abstract: Optimizing the expected values of probabilistic processes is a central problem in computer science and its applications, arising in fields ranging from artificial intelligence to operations research to statistical computing. Unfortunately, automatic differentiation techniques developed for deterministic programs do not in general compute the correct gradients needed for widely used solutions based… ▽ More Optimizing the expected values of probabilistic processes is a central problem in computer science and its applications, arising in fields ranging from artificial intelligence to operations research to statistical computing. Unfortunately, automatic differentiation techniques developed for deterministic programs do not in general compute the correct gradients needed for widely used solutions based on gradient-based optimization. In this paper, we present ADEV, an extension to forward-mode AD that correctly differentiates the expectations of probabilistic processes represented as programs that make random choices. Our algorithm is a source-to-source program transformation on an expressive, higher-order language for probabilistic computation, with both discrete and continuous probability distributions. The result of our transformation is a new probabilistic program, whose expected return value is the derivative of the original program's expectation. This output program can be run to generate unbiased Monte Carlo estimates of the desired gradient, which can then be used within the inner loop of stochastic gradient descent. We prove ADEV correct using logical relations over the denotations of the source and target probabilistic programs. Because it modularly extends forward-mode AD, our algorithm lends itself to a concise implementation strategy, which we exploit to develop a prototype in just a few dozen lines of Haskell (https://github.com/probcomp/adev). △ Less

Submitted 13 December, 2022; originally announced December 2022.

Comments: to appear at POPL 2023

Journal ref: POPL 2023

Architected Materials for Mechanical Compression: Design via Simulation, Deep Learning, and Experimentation

Authors: Andrew J. Lew, Kai **, Markus J. Buehler

Abstract: Architected materials can achieve enhanced properties compared to their plain counterparts. Specific architecting serves as a powerful design lever to achieve targeted behavior without changing the base material. Thus, the connection between architected structure and resultant properties remains an open field of great interest to many fields, from aerospace to civil to automotive applications. Her… ▽ More Architected materials can achieve enhanced properties compared to their plain counterparts. Specific architecting serves as a powerful design lever to achieve targeted behavior without changing the base material. Thus, the connection between architected structure and resultant properties remains an open field of great interest to many fields, from aerospace to civil to automotive applications. Here, we focus on properties related to mechanical compression, and design hierarchical honeycomb structures to meet specific values of stiffness and compressive stress. To do so, we employ a combination of techniques in a singular workflow, starting with molecular dynamics simulation of the forward design problem, augmenting with data-driven artificial intelligence models to address the inverse design problem, and verifying the behavior of de novo structures with experimentation of additively manufactured samples. We thereby demonstrate an approach for architected design that is generalizable to multiple material properties and agnostic to the identity of the base material. △ Less

Submitted 13 February, 2023; v1 submitted 5 December, 2022; originally announced December 2022.

Physics-Based Nozzle Design Rules for High-Frequency Liquid Metal Jetting

Authors: Jongmin Seo, Christoforos Somarakis, Svyatoslav Korneev, Morad Behandish, Adrian J. Lew

Abstract: We present physics-based nozzle design rules to achieve high-throughput and stable jetting in drop-on-demand liquid metal 3D printing. The design rules are based on scaling laws that capture the change of meniscus oscillation relaxation time with geometric characteristics of the nozzle's inner profile. These characteristics include volume, cross-sectional area, and inner surface area of the nozzle… ▽ More We present physics-based nozzle design rules to achieve high-throughput and stable jetting in drop-on-demand liquid metal 3D printing. The design rules are based on scaling laws that capture the change of meniscus oscillation relaxation time with geometric characteristics of the nozzle's inner profile. These characteristics include volume, cross-sectional area, and inner surface area of the nozzle. Using boundary layer theory for a simple geometry, we show that the meniscus settles faster when the ratio of inner surface area to volume is increased. High-fidelity multiphase flow simulations verify this scaling. We use these laws to explore several design concepts with parameterized classes of shapes that reduce the meniscus relaxation time while preserving desired droplet specs. Finally, we show that for various nozzle profile concepts, the optimal performance can be achieved by increasing the ratio of the circumferential surface area to the bulk volume to the extent that is allowable by manufacturing constraints. △ Less

Submitted 14 July, 2022; originally announced July 2022.

Comments: Under Review in Physics of Fluids, AIP Publishing

On the $d$-dimensional algebraic connectivity of graphs

Authors: Alan Lew, Eran Nevo, Yuval Peled, Orit E. Raz

Abstract: The $d$-dimensional algebraic connectivity $a_d(G)$ of a graph $G=(V,E)$, introduced by Jordán and Tanigawa, is a quantitative measure of the $d$-dimensional rigidity of $G$ that is defined in terms of the eigenvalues of stiffness matrices (which are analogues of the graph Laplacian) associated to map**s of the vertex set $V$ into $\mathbb{R}^d$. Here, we analyze the $d$-dimensional algebraic… ▽ More The $d$-dimensional algebraic connectivity $a_d(G)$ of a graph $G=(V,E)$, introduced by Jordán and Tanigawa, is a quantitative measure of the $d$-dimensional rigidity of $G$ that is defined in terms of the eigenvalues of stiffness matrices (which are analogues of the graph Laplacian) associated to map**s of the vertex set $V$ into $\mathbb{R}^d$. Here, we analyze the $d$-dimensional algebraic connectivity of complete graphs. In particular, we show that, for $d\geq 3$, $a_d(K_{d+1})=1$, and for $n\geq 2d$, \[ \left\lceil\frac{n}{2d}\right\rceil-2d+1\leq a_d(K_n) \leq \frac{2n}{3(d-1)}+\frac{1}{3}. \] △ Less

Submitted 11 May, 2022; originally announced May 2022.

Recursive Monte Carlo and Variational Inference with Auxiliary Variables

Authors: Alexander K. Lew, Marco Cusumano-Towner, Vikash K. Mansinghka

Abstract: A key design constraint when implementing Monte Carlo and variational inference algorithms is that it must be possible to cheaply and exactly evaluate the marginal densities of proposal distributions and variational families. This takes many interesting proposals off the table, such as those based on involved simulations or stochastic optimization. This paper broadens the design space, by presenti… ▽ More A key design constraint when implementing Monte Carlo and variational inference algorithms is that it must be possible to cheaply and exactly evaluate the marginal densities of proposal distributions and variational families. This takes many interesting proposals off the table, such as those based on involved simulations or stochastic optimization. This paper broadens the design space, by presenting a framework for applying Monte Carlo and variational inference algorithms when proposal densities cannot be exactly evaluated. Our framework, recursive auxiliary-variable inference (RAVI), instead approximates the necessary densities using meta-inference: an additional layer of Monte Carlo or variational inference, that targets the proposal, rather than the model. RAVI generalizes and unifies several existing methods for inference with expressive approximating families, which we show correspond to specific choices of meta-inference algorithm, and provides new theory for analyzing their bias and variance. We illustrate RAVI's design framework and theorems by using them to analyze and improve upon Salimans et al.'s Markov Chain Variational Inference, and to design a novel sampler for Dirichlet process mixtures, achieving state-of-the-art results on a standard benchmark dataset from astronomy and on a challenging datacleaning task with Medicare hospital data. △ Less

Submitted 18 November, 2022; v1 submitted 5 March, 2022; originally announced March 2022.

Comments: version published at UAI

Journal ref: UAI 2022

Sharp threshold for rigidity of random graphs

Authors: Alan Lew, Eran Nevo, Yuval Peled, Orit E. Raz

Abstract: We consider the Erdős-Rényi evolution of random graphs, where a new uniformly distributed edge is added to the graph in every step. For every fixed $d\ge 1$, we show that with high probability, the graph becomes rigid in $\mathbb R^d$ at the very moment its minimum degree becomes $d$, and it becomes globally rigid in $\mathbb R^d$ at the very moment its minimum degree becomes $d+1$. We consider the Erdős-Rényi evolution of random graphs, where a new uniformly distributed edge is added to the graph in every step. For every fixed $d\ge 1$, we show that with high probability, the graph becomes rigid in $\mathbb R^d$ at the very moment its minimum degree becomes $d$, and it becomes globally rigid in $\mathbb R^d$ at the very moment its minimum degree becomes $d+1$. △ Less

Submitted 13 September, 2022; v1 submitted 20 February, 2022; originally announced February 2022.

MSC Class: 05C80; 52C25

Towards Denotational Semantics of AD for Higher-Order, Recursive, Probabilistic Languages

Authors: Alexander K. Lew, Mathieu Huot, Vikash K. Mansinghka

Abstract: Automatic differentiation (AD) aims to compute derivatives of user-defined functions, but in Turing-complete languages, this simple specification does not fully capture AD's behavior: AD sometimes disagrees with the true derivative of a differentiable program, and when AD is applied to non-differentiable or effectful programs, it is unclear what guarantees (if any) hold of the resulting code. We s… ▽ More Automatic differentiation (AD) aims to compute derivatives of user-defined functions, but in Turing-complete languages, this simple specification does not fully capture AD's behavior: AD sometimes disagrees with the true derivative of a differentiable program, and when AD is applied to non-differentiable or effectful programs, it is unclear what guarantees (if any) hold of the resulting code. We study an expressive differentiable programming language, with piecewise-analytic primitives, higher-order functions, and general recursion. Our main result is that even in this general setting, a version of Lee et al. [2020]'s correctness theorem (originally proven for a first-order language without partiality or recursion) holds: all programs denote so-called $ω$PAP functions, and AD computes correct intensional derivatives of them. Mazza and Pagani [2021]'s recent theorem, that AD disagrees with the true derivative of a differentiable recursive program at a measure-zero set of inputs, can be derived as a straightforward corollary of this fact. We also apply the framework to study probabilistic programs, and recover a recent result from Mak et al. [2021] via a novel denotational argument. △ Less

Submitted 6 December, 2021; v1 submitted 30 November, 2021; originally announced November 2021.

Comments: Presented at the NeurIPS 2021 differentiable programming workshop

Leray numbers of tolerance complexes

Authors: Minki Kim, Alan Lew

Abstract: Let $K$ be a simplicial complex on vertex set $V$. $K$ is called $d$-Leray if the homology groups of any induced subcomplex of $K$ are trivial in dimensions $d$ and higher. $K$ is called $d$-collapsible if it can be reduced to the void complex by sequentially removing a simplex of size at most $d$ that is contained in a unique maximal face. We define the $t$-tolerance complex of $K$,… ▽ More Let $K$ be a simplicial complex on vertex set $V$. $K$ is called $d$-Leray if the homology groups of any induced subcomplex of $K$ are trivial in dimensions $d$ and higher. $K$ is called $d$-collapsible if it can be reduced to the void complex by sequentially removing a simplex of size at most $d$ that is contained in a unique maximal face. We define the $t$-tolerance complex of $K$, $\mathcal{T}_t(K)$, as the simplicial complex on vertex set $V$ whose simplices are formed as the union of a simplex in $K$ and a set of size at most $t$. We prove that for any $d$ and $t$ there exists a positive integer $h(t,d)$ such that, for every $d$-collapsible complex $K$, the $t$-tolerance complex $\mathcal{T}_t(K)$ is $h(t,d)$-Leray. The definition of the complex $\mathcal{T}_t(K)$ is motivated by results of Montejano and Oliveros on "tolerant" versions of Helly's theorem. As an application, we present some new tolerant versions of the colorful Helly theorem. △ Less

Submitted 7 September, 2021; originally announced September 2021.

Small-misorientation toughness in biominerals evolved convergently

Authors: Andrew J. Lew, Cayla A. Stifler, Connor A. Schmidt, Markus J. Buehler, Pupa U. P. A. Gilbert

Abstract: The hardest materials in living organisms are biologically grown crystalline minerals, or biominerals, which are also incredibly fracture-tough. Biomineral mesostructure includes size, shape, spatial arrangement, and crystal orientation of crystallites, observable at the mesoscale (10 nanometer - 10 micron). Here we show that diverse biominerals, including nacre and prisms from mollusk shells, cor… ▽ More The hardest materials in living organisms are biologically grown crystalline minerals, or biominerals, which are also incredibly fracture-tough. Biomineral mesostructure includes size, shape, spatial arrangement, and crystal orientation of crystallites, observable at the mesoscale (10 nanometer - 10 micron). Here we show that diverse biominerals, including nacre and prisms from mollusk shells, coral skeletons, and tunicate spicules have different mesostructures, but they converged to similar, small (<30 degrees) misorientations of adjacent crystals at the mesoscale. We show that such small misorientations are an effective toughening mechanism. Combining Polarization-dependent Imaging Contrast (PIC) map** of mesostructures and Molecular Dynamics (MD) simulations of misoriented bicrystals, we reveal here that small misorientations toughen bicrystals, thus explaining why they evolved independently but convergently: preventing fracture is a clear evolutionary advantage for diverse organisms. △ Less

Submitted 17 August, 2021; originally announced August 2021.

Comments: 18 pages, 6 figures

Representability and boxicity of simplicial complexes

Authors: Alan Lew

Abstract: Let $X$ be a simplicial complex on vertex set $V$. We say that $X$ is $d$-representable if it is isomorphic to the nerve of a family of convex sets in $\mathbb{R}^d$. We define the $d$-boxicity of $X$ as the minimal $k$ such that $X$ can be written as the intersection of $k$ $d$-representable simplicial complexes. This generalizes the notion of boxicity of a graph, defined by Roberts. A missing… ▽ More Let $X$ be a simplicial complex on vertex set $V$. We say that $X$ is $d$-representable if it is isomorphic to the nerve of a family of convex sets in $\mathbb{R}^d$. We define the $d$-boxicity of $X$ as the minimal $k$ such that $X$ can be written as the intersection of $k$ $d$-representable simplicial complexes. This generalizes the notion of boxicity of a graph, defined by Roberts. A missing face of $X$ is a set $τ\subset V$ such that $τ\notin X$ but $σ\in X$ for any $σ\subsetneq τ$. We prove that the $d$-boxicity of a simplicial complex on $n$ vertices without missing faces of dimension larger than $d$ is at most $\left\lfloor\frac{1}{d+1}\binom{n}{d}\right\rfloor$. The bound is sharp: the $d$-boxicity of a simplicial complex whose set of missing faces form a Steiner $(d,d+1,n)$-system is exactly $\frac{1}{d+1}\binom{n}{d}$. △ Less

Submitted 23 August, 2020; originally announced August 2020.

PClean: Bayesian Data Cleaning at Scale with Domain-Specific Probabilistic Programming

Authors: Alexander K. Lew, Monica Agrawal, David Sontag, Vikash K. Mansinghka

Abstract: Data cleaning is naturally framed as probabilistic inference in a generative model of ground-truth data and likely errors, but the diversity of real-world error patterns and the hardness of inference make Bayesian approaches difficult to automate. We present PClean, a probabilistic programming language (PPL) for leveraging dataset-specific knowledge to automate Bayesian cleaning. Compared to gener… ▽ More Data cleaning is naturally framed as probabilistic inference in a generative model of ground-truth data and likely errors, but the diversity of real-world error patterns and the hardness of inference make Bayesian approaches difficult to automate. We present PClean, a probabilistic programming language (PPL) for leveraging dataset-specific knowledge to automate Bayesian cleaning. Compared to general-purpose PPLs, PClean tackles a restricted problem domain, enabling three modeling and inference innovations: (1) a non-parametric model of relational database instances, which users' programs customize; (2) a novel sequential Monte Carlo inference algorithm that exploits the structure of PClean's model class; and (3) a compiler that generates near-optimal SMC proposals and blocked-Gibbs rejuvenation kernels based on the user's model and data. We show empirically that short (< 50-line) PClean programs can: be faster and more accurate than generic PPL inference on data-cleaning benchmarks; match state-of-the-art data-cleaning systems in terms of accuracy and runtime (unlike generic PPL inference in the same runtime); and scale to real-world datasets with millions of records. △ Less

Submitted 18 November, 2022; v1 submitted 23 July, 2020; originally announced July 2020.

Comments: Published version

Journal ref: AISTATS 2021

Automating Involutive MCMC using Probabilistic and Differentiable Programming

Authors: Marco Cusumano-Towner, Alexander K. Lew, Vikash K. Mansinghka

Abstract: Involutive MCMC is a unifying mathematical construction for MCMC kernels that generalizes many classic and state-of-the-art MCMC algorithms, from reversible jump MCMC to kernels based on deep neural networks. But as with MCMC samplers more generally, implementing involutive MCMC kernels is often tedious and error-prone, especially when sampling on complex state spaces. This paper describes a techn… ▽ More Involutive MCMC is a unifying mathematical construction for MCMC kernels that generalizes many classic and state-of-the-art MCMC algorithms, from reversible jump MCMC to kernels based on deep neural networks. But as with MCMC samplers more generally, implementing involutive MCMC kernels is often tedious and error-prone, especially when sampling on complex state spaces. This paper describes a technique for automating the implementation of involutive MCMC kernels given (i) a pair of probabilistic programs defining the target distribution and an auxiliary distribution respectively and (ii) a differentiable program that transforms the execution traces of these probabilistic programs. The technique, which is implemented as part of the Gen probabilistic programming system, also automatically detects user errors in the specification of involutive MCMC kernels and exploits sparsity in the kernels for improved efficiency. The paper shows example Gen code for a split-merge reversible jump move in an infinite Gaussian mixture model and a state-dependent mixture of proposals on a combinatorial space of covariance functions for a Gaussian process. △ Less

Submitted 20 July, 2020; v1 submitted 19 July, 2020; originally announced July 2020.

Spring-Slider and Finite Element Modeling of Microseismic Events and Fault Slip during Hydraulic Fracturing

Authors: Ali Kashefi, Eric M. Dunham, Benjamin Grossman-Ponemon, Adrian J. Lew

Abstract: Hydraulic fracturing increases reservoir permeability by opening fractures and triggering slip on natural fractures and faults. While seismic slip of small faults or fault patches is detectable as microseismic events, the role of aseismic slip is poorly understood. From a modeling standpoint, geomechanical analysis using the Coulomb criterion can determine if faults slip but not whether slip is se… ▽ More Hydraulic fracturing increases reservoir permeability by opening fractures and triggering slip on natural fractures and faults. While seismic slip of small faults or fault patches is detectable as microseismic events, the role of aseismic slip is poorly understood. From a modeling standpoint, geomechanical analysis using the Coulomb criterion can determine if faults slip but not whether slip is seismic or aseismic. Here we propose a computational methodology to predict fault slip, and whether slip is seismic or aseismic, using rate-and-state friction. To avoid computational costs associated with resolving small faults, we use the spring-slider idealization that treats faults as points. Interaction between faults is neglected. The method is applied to study fault slip from a hydraulic fracture that grows past a fault, without intersecting it. We represent the hydraulic fracture stressing using an asymptotic expansion of stresses around the tip of a tensile crack. We investigate the effect of fault length, orientation, and distance from the hydraulic fracture. For velocity-weakening faults with stiffness smaller than a critical stiffness, slip is seismic, whereas faults with stiffness greater than the critical stiffness slip aseismically. Furthermore, we compare the spring-slider idealization with a finite element analysis that resolves spatially variable slip. The spring-slider idealization provides reasonably accurate predictions of moment and even moment-rate history, especially for faults having stiffness close to or larger than the critical stiffness. Differences appear for large faults where rupture propagation is important, though differences might still be negligible for many applications. △ Less

Submitted 19 February, 2020; originally announced February 2020.

Complexes of graphs with bounded independence number

Authors: Minki Kim, Alan Lew

Abstract: Let $G=(V,E)$ be a graph and $n$ a positive integer. Let $I_n(G)$ be the abstract simplicial complex whose simplices are the subsets of $V$ that do not contain an independent set of size $n$ in $G$. We study the collapsibility numbers of the complexes $I_n(G)$ for various classes of graphs, focusing on the class of graphs with maximum degree bounded by $Δ$. As an application, we obtain the followi… ▽ More Let $G=(V,E)$ be a graph and $n$ a positive integer. Let $I_n(G)$ be the abstract simplicial complex whose simplices are the subsets of $V$ that do not contain an independent set of size $n$ in $G$. We study the collapsibility numbers of the complexes $I_n(G)$ for various classes of graphs, focusing on the class of graphs with maximum degree bounded by $Δ$. As an application, we obtain the following result: Let $G$ be a claw-free graph with maximum degree at most $Δ$. Then, every collection of $\left\lfloor\left(\fracΔ{2}+1\right)(n-1)\right\rfloor+1$ independent sets in $G$ has a rainbow independent set of size $n$. △ Less

Submitted 29 December, 2019; originally announced December 2019.

Bayesian causal inference via probabilistic program synthesis

Authors: Sam Witty, Alexander Lew, David Jensen, Vikash Mansinghka

Abstract: Causal inference can be formalized as Bayesian inference that combines a prior distribution over causal models and likelihoods that account for both observations and interventions. We show that it is possible to implement this approach using a sufficiently expressive probabilistic programming language. Priors are represented using probabilistic programs that generate source code in a domain specif… ▽ More Causal inference can be formalized as Bayesian inference that combines a prior distribution over causal models and likelihoods that account for both observations and interventions. We show that it is possible to implement this approach using a sufficiently expressive probabilistic programming language. Priors are represented using probabilistic programs that generate source code in a domain specific language. Interventions are represented using probabilistic programs that edit this source code to modify the original generative process. This approach makes it straightforward to incorporate data from atomic interventions, as well as shift interventions, variance-scaling interventions, and other interventions that modify causal structure. This approach also enables the use of general-purpose inference machinery for probabilistic programs to infer probable causal structures and parameters from data. This abstract describes a prototype of this approach in the Gen probabilistic programming language. △ Less

Submitted 30 October, 2019; originally announced October 2019.

Few-Shot Bayesian Imitation Learning with Logical Program Policies

Authors: Tom Silver, Kelsey R. Allen, Alex K. Lew, Leslie Pack Kaelbling, Josh Tenenbaum

Abstract: Humans can learn many novel tasks from a very small number (1--5) of demonstrations, in stark contrast to the data requirements of nearly tabula rasa deep learning methods. We propose an expressive class of policies, a strong but general prior, and a learning algorithm that, together, can learn interesting policies from very few examples. We represent policies as logical combinations of programs d… ▽ More Humans can learn many novel tasks from a very small number (1--5) of demonstrations, in stark contrast to the data requirements of nearly tabula rasa deep learning methods. We propose an expressive class of policies, a strong but general prior, and a learning algorithm that, together, can learn interesting policies from very few examples. We represent policies as logical combinations of programs drawn from a domain-specific language (DSL), define a prior over policies with a probabilistic grammar, and derive an approximate Bayesian inference algorithm to learn policies from demonstrations. In experiments, we study five strategy games played on a 2D grid with one shared DSL. After a few demonstrations of each game, the inferred policies generalize to new game instances that differ substantially from the demonstrations. Our policy learning is 20--1,000x more data efficient than convolutional and fully convolutional policy learning and many orders of magnitude more computationally efficient than vanilla program induction. We argue that the proposed method is an apt choice for tasks that have scarce training data and feature significant, structured variation between task instances. △ Less

Submitted 16 November, 2019; v1 submitted 12 April, 2019; originally announced April 2019.

Comments: AAAI 2020

Collapsibility of simplicial complexes of hypergraphs

Authors: Alan Lew

Abstract: Let $\mathcal{H}$ be a hypergraph of rank $r$. We show that the simplicial complex whose simplices are the hypergraphs $\mathcal{F}\subset\mathcal{H}$ with covering number at most $p$ is $\left(\binom{r+p}{r}-1\right)$-collapsible, and the simplicial complex whose simplices are the pairwise intersecting hypergraphs $\mathcal{F}\subset\mathcal{H}$ is $\frac{1}{2}\binom{2r}{r}$-collapsible. Let $\mathcal{H}$ be a hypergraph of rank $r$. We show that the simplicial complex whose simplices are the hypergraphs $\mathcal{F}\subset\mathcal{H}$ with covering number at most $p$ is $\left(\binom{r+p}{r}-1\right)$-collapsible, and the simplicial complex whose simplices are the pairwise intersecting hypergraphs $\mathcal{F}\subset\mathcal{H}$ is $\frac{1}{2}\binom{2r}{r}$-collapsible. △ Less

Submitted 14 October, 2019; v1 submitted 28 October, 2018; originally announced October 2018.

Journal ref: The Electronic Journal of Combinatorics, Volume 26, Issue 4 (2019), #P4.10

Spectral gaps, missing faces and minimal degrees

Authors: Alan Lew

Abstract: Let $X$ be a simplicial complex with $n$ vertices. A missing face of $X$ is a simplex $σ\notin X$ such that $τ\in X$ for any $τ\subsetneq σ$. For a $k$-dimensional simplex $σ$ in $X$, its degree in $X$ is the number of $(k+1)$-dimensional simplices in $X$ containing it. Let $δ_k$ denote the minimal degree of a $k$-dimensional simplex in $X$. Let $L_k$ denote the $k$-Laplacian acting on real $k$-co… ▽ More Let $X$ be a simplicial complex with $n$ vertices. A missing face of $X$ is a simplex $σ\notin X$ such that $τ\in X$ for any $τ\subsetneq σ$. For a $k$-dimensional simplex $σ$ in $X$, its degree in $X$ is the number of $(k+1)$-dimensional simplices in $X$ containing it. Let $δ_k$ denote the minimal degree of a $k$-dimensional simplex in $X$. Let $L_k$ denote the $k$-Laplacian acting on real $k$-cochains of $X$ and let $μ_k(X)$ denote its minimal eigenvalue. We prove the following lower bound on the spectral gaps $μ_k(X)$, for complexes $X$ without missing faces of dimension larger than $d$: \[ μ_k(X)\geq (d+1)(δ_k+k+1)-d n. \] As a consequence we obtain a new proof of a vanishing result for the homology of simplicial complexes without large missing faces. We present a family of examples achieving equality at all dimensions, showing that the bound is tight. For $d=1$ we characterize the equality case. △ Less

Submitted 4 July, 2018; originally announced July 2018.

Journal ref: Journal of Combinatorial Theory, Series A 169 (2020) 105127

Spectral gaps of simplicial complexes without large missing faces

Authors: Alan Lew

Abstract: Let $X$ be a simplicial complex on $n$ vertices without missing faces of dimension larger than $d$. Let $L_{j}$ denote the $j$-Laplacian acting on real $j$-cochains of $X$ and let $μ_{j}(X)$ denote its minimal eigenvalue. We study the connection between the spectral gaps $μ_{k}(X)$ for $k\geq d$ and $μ_{d-1}(X)$. In particular, we establish the following vanishing result: If… ▽ More Let $X$ be a simplicial complex on $n$ vertices without missing faces of dimension larger than $d$. Let $L_{j}$ denote the $j$-Laplacian acting on real $j$-cochains of $X$ and let $μ_{j}(X)$ denote its minimal eigenvalue. We study the connection between the spectral gaps $μ_{k}(X)$ for $k\geq d$ and $μ_{d-1}(X)$. In particular, we establish the following vanishing result: If $μ_{d-1}(X)>(1-\binom{k+1}{d}^{-1})n$, then $\tilde{H}^{j}(X;\mathbb{R})=0$ for all $d-1\leq j \leq k$. As an application we prove a fractional extension of a Hall-type theorem of Holmsen, Martínez-Sandoval and Montejano for general position sets in matroids. △ Less

Submitted 1 June, 2017; originally announced June 2017.

Journal ref: International Mathematics Research Notices, rny115 (2018)

Universal Meshes for the Simulation of Brittle Fracture and Moving Boundary Problems

Authors: Maurizio M. Chiaramonte, Evan S. Gawlik, Hardik Kabaria, Adrian J. Lew

Abstract: Universal meshes have recently appeared in the literature as a compu- tationally efficient and robust paradigm for the generation of conforming simpli- cial meshes for domains with evolving boundaries. The main idea behind a univer- sal mesh is to immerse the moving boundary in a background mesh (the universal mesh), and to produce a mesh that conforms to the moving boundary at any given time by a… ▽ More Universal meshes have recently appeared in the literature as a compu- tationally efficient and robust paradigm for the generation of conforming simpli- cial meshes for domains with evolving boundaries. The main idea behind a univer- sal mesh is to immerse the moving boundary in a background mesh (the universal mesh), and to produce a mesh that conforms to the moving boundary at any given time by adjusting a few of elements of the background mesh. In this manuscript we present the application of universal meshes to the simulation of brittle fracturing. To this extent, we provide a high level description of a crack propagation algorithm and showcase its capabilities. Alongside universal meshes for the simulation of brit- tle fracture, we provide other examples for which universal meshes prove to be a powerful tool, namely fluid flow past moving obstacles. Lastly, we conclude the manuscript with some remarks on the current state of universal meshes and future directions. △ Less

Submitted 17 October, 2015; originally announced October 2015.

Computing stress intensity factors for curvilinear cracks

Authors: Maurizio M. Chiaramonte, Yongxing Shen, Leon M. Keer, Adrian J. Lew

Abstract: The use of the interaction integral to compute stress intensity factors around a crack tip requires selecting an auxiliary field and a material variation field. We formulate a family of these fields accounting for the curvilinear nature of cracks that, in conjunction with a discrete formulation of the interaction integral, yield optimally convergent stress intensity factors. We formulate three pai… ▽ More The use of the interaction integral to compute stress intensity factors around a crack tip requires selecting an auxiliary field and a material variation field. We formulate a family of these fields accounting for the curvilinear nature of cracks that, in conjunction with a discrete formulation of the interaction integral, yield optimally convergent stress intensity factors. We formulate three pairs of auxiliary and material variation fields chosen to yield a simple expression of the interaction integral for different classes of problems. The formulation accounts for crack face tractions and body forces. Distinct features of the fields are their ease of construction and implementation. The resulting stress intensity factors are observed converging at a rate that doubles the one of the stress field. We provide a sketch of the theoretical justification for the observed convergence rates, and discuss issues such as quadratures and domain approximations needed to attain such convergent behavior. Through two representative examples, a circular arc crack and a loaded power function crack, we illustrate the convergence rates of the computed stress intensity factors. The numerical results also show the independence of the method on the size of the domain of integration. △ Less

Submitted 13 January, 2015; originally announced January 2015.

Supercloseness of Orthogonal Projections onto Nearby Finite Element Spaces

Authors: Evan S. Gawlik, Adrian J. Lew

Abstract: We derive upper bounds on the difference between the orthogonal projections of a smooth function $u$ onto two finite element spaces that are nearby, in the sense that the support of every shape function belonging to one but not both of the spaces is contained in a common region whose measure tends to zero under mesh refinement. The bounds apply, in particular, to the setting in which the two finit… ▽ More We derive upper bounds on the difference between the orthogonal projections of a smooth function $u$ onto two finite element spaces that are nearby, in the sense that the support of every shape function belonging to one but not both of the spaces is contained in a common region whose measure tends to zero under mesh refinement. The bounds apply, in particular, to the setting in which the two finite element spaces consist of continuous functions that are elementwise polynomials over shape-regular, quasi-uniform meshes that coincide except on a region of measure $O(h^γ)$, where $γ$ is a nonnegative scalar and $h$ is the mesh spacing. The projector may be, for example, the orthogonal projector with respect to the $L^2$- or $H^1$-inner product. In these and other circumstances, the bounds are superconvergent under a few mild regularity assumptions. That is, under mesh refinement, the two projections differ in norm by an amount that decays to zero at a faster rate than the amounts by which each projection differs from $u$. We present numerical examples to illustrate these superconvergent estimates and verify the necessity of the regularity assumptions on $u$. △ Less

Submitted 18 August, 2014; originally announced August 2014.

MSC Class: 65N30 (Primary); 65N15 (Secondary)

High-Order Finite Element Methods for Moving Boundary Problems with Prescribed Boundary Evolution

Authors: Evan S. Gawlik, Adrian J. Lew

Abstract: We introduce a framework for the design of finite element methods for two-dimensional moving boundary problems with prescribed boundary evolution that have arbitrarily high order of accuracy, both in space and in time. At the core of our approach is the use of a universal mesh: a stationary background mesh containing the domain of interest for all times that adapts to the geometry of the immersed… ▽ More We introduce a framework for the design of finite element methods for two-dimensional moving boundary problems with prescribed boundary evolution that have arbitrarily high order of accuracy, both in space and in time. At the core of our approach is the use of a universal mesh: a stationary background mesh containing the domain of interest for all times that adapts to the geometry of the immersed domain by adjusting a small number of mesh elements in the neighborhood of the moving boundary. The resulting method maintains an exact representation of the (prescribed) moving boundary at the discrete level, or an approximation of the appropriate order, yet is immune to large distortions of the mesh under large deformations of the domain. The framework is general, making it possible to achieve any desired order of accuracy in space and time by selecting a preferred and suitable finite-element space on the universal mesh for the problem at hand, and a preferred and suitable time integrator for ordinary differential equations. We illustrate our approach by constructing a particular class of methods, and apply them to a prescribed-boundary variant of the Stefan problem. We present numerical evidence for the order of accuracy of our schemes in one and two dimensions. △ Less

Submitted 8 May, 2014; originally announced May 2014.

MSC Class: 65M60 (Primary); 65M50 (Secondary)

Variational integrators for the dynamics of thermo-elastic solids with finite speed thermal waves

Authors: Pablo Mata A, Adrian J Lew

Abstract: This paper formulates variational integrators for finite element discretizations of deformable bodies with heat conduction in the form of finite speed thermal waves. The cornerstone of the construction consists in taking advantage of the fact that the Green-Naghdi theory of type II for thermo-elastic solids has a Hamiltonian structure. Thus, standard techniques to construct variational integrators… ▽ More This paper formulates variational integrators for finite element discretizations of deformable bodies with heat conduction in the form of finite speed thermal waves. The cornerstone of the construction consists in taking advantage of the fact that the Green-Naghdi theory of type II for thermo-elastic solids has a Hamiltonian structure. Thus, standard techniques to construct variational integrators can be applied to finite element discretizations of the problem. The resulting discrete-in-time trajectories are then consistent with the laws of thermodynamics for these systems: for an isolated system, they exactly conserve the total entropy, and nearly exactly conserve the total energy over exponentially long periods of time. Moreover, linear and angular momenta are also exactly conserved whenever the exact system does. For definiteness, we construct an explicit second-order accurate algorithm for affine tetrahedral elements in two and three-dimensions, and demonstrate its performance with numerical examples. △ Less

Submitted 14 March, 2014; originally announced March 2014.

Journal ref: J.Comp.Phys. 257 (2014) 1423-1443

Universal Meshes: A new paradigm for computing with nonconforming triangulations

Authors: Ramsharan Rangarajan, Adrian J. Lew

Abstract: We describe a method for discretizing planar C2-regular domains immersed in non-conforming triangulations. The method consists in constructing map**s from triangles in a background mesh to curvilinear ones that conform exactly to the immersed domain. Constructing such a map relies on a novel way of parameterizing the immersed boundary over a collection of nearby edges with its closest point proj… ▽ More We describe a method for discretizing planar C2-regular domains immersed in non-conforming triangulations. The method consists in constructing map**s from triangles in a background mesh to curvilinear ones that conform exactly to the immersed domain. Constructing such a map relies on a novel way of parameterizing the immersed boundary over a collection of nearby edges with its closest point projection. By interpolating the map**s to curvilinear triangles at select points, we recover isoparametric map**s for the immersed domain defined over the background mesh. Indeed, interpolating the constructed map**s just at the vertices of the background mesh yields a fast meshing algorithm that involves only perturbing a few vertices near the boundary. For the discretization of a curved domain to be robust, we have to impose restrictions on the background mesh. Conversely, these restrictions define a family of domains that can be discretized with a given background mesh. We then say that the background mesh is a universal mesh for such a family of domains. The notion of universal meshes is particularly useful in free/moving boundary problems because the same background mesh can serve as the universal mesh for the evolving domain for time intervals that are independent of the time step. Hence it facilitates a framework for finite element calculations over evolving domains while using a fixed background mesh. Furthermore, since the evolving geometry can be approximated with any desired order, numerical solutions can be computed with high-order accuracy. We demonstrate these ideas with various numerical examples. △ Less

Submitted 23 January, 2012; originally announced January 2012.

MSC Class: 65N30; 68U05; 65M50; 65N50

Analysis of a method to parameterize planar curves immersed in triangulations

Authors: Ramsharan Rangarajan, Adrian J. Lew

Abstract: We prove that a planar $C^2$-regular boundary $Γ$ can always be parameterized with its closest point projection $π$ over a certain collection of edges $Γ_h$ in an ambient triangulation, by making simple assumptions on the background mesh. For $Γ_h$, we select the edges that have both vertices on one side of $Γ$ and belong to a triangle that has a vertex on the other side. By imposing restrictions… ▽ More We prove that a planar $C^2$-regular boundary $Γ$ can always be parameterized with its closest point projection $π$ over a certain collection of edges $Γ_h$ in an ambient triangulation, by making simple assumptions on the background mesh. For $Γ_h$, we select the edges that have both vertices on one side of $Γ$ and belong to a triangle that has a vertex on the other side. By imposing restrictions on the size of triangles near the curve and by requesting that certain angles in the mesh be strictly acute, we prove that $π:Γ_h\rightarrowΓ$ is a homeomorphism, that it is $C^1$ on each edge in $Γ_h$ and provide bounds for the Jacobian of the parameterization. The assumptions on the background mesh are both easy to satisfy in practice and conveniently verified in computer implementations. The parameterization analyzed here was previously proposed by the authors and applied to the construction of high-order curved finite elements on a class of planar piecewise $C^2$-curves. △ Less

Submitted 16 February, 2013; v1 submitted 16 September, 2011; originally announced September 2011.

MSC Class: 68U05; 65D18

A microstructurally informed model for the mechanical response of three-dimensional actin networks

Authors: Ronald Y. Kwon, Adrian J. Lew, Christopher R. Jacobs

Abstract: We propose a class of microstructurally informed models for the linear elastic mechanical behavior of cross-linked polymer networks such as the actin cytoskeleton. Salient features of the models include the possibility to represent anisotropic mechanical behavior resulting from anisotropic filament distributions, and a power-law scaling of the mechanical properties with the filament density. Mec… ▽ More We propose a class of microstructurally informed models for the linear elastic mechanical behavior of cross-linked polymer networks such as the actin cytoskeleton. Salient features of the models include the possibility to represent anisotropic mechanical behavior resulting from anisotropic filament distributions, and a power-law scaling of the mechanical properties with the filament density. Mechanical models within the class are parameterized by seven different constants. We demonstrate a procedure for determining these constants using finite element models of three-dimensional actin networks. Actin filaments and cross-links were modeled as elastic rods, and the networks were constructed at physiological volume fractions and at the scale of an image voxel. We show the performance of the model in estimating the mechanical behavior of the networks over a wide range of filament densities and degrees of anisotropy. △ Less

Submitted 26 February, 2008; originally announced February 2008.

Comments: 26 pages, 9 figures