Publications

Parametric computer-aided design (CAD) tools are the predominant way that engineers specify physical structures, from bicycle pedals to airplanes to printed circuit boards. The key characteristic of parametric CAD is that design intent is encoded not only via geometric primitives, but also by parameterized constraints between the elements. This relational specification can be viewed as the construction of a constraint program, allowing edits to coherently propagate to other parts of the design. Machine learning offers the intriguing possibility of accelerating the design process via generative modeling of these structures, enabling new tools such as autocompletion, constraint inference, and conditional synthesis. In this work, we present such an approach to generative modeling of parametric CAD sketches, which constitute the basic computational building blocks of modern mechanical design. Our model, trained on real-world designs from the SketchGraphs dataset, autoregressively synthesizes sketches as sequences of primitives, with initial coordinates, and constraints that reference back to the sampled primitives. As samples from the model match the constraint graph representation used in standard CAD software, they may be directly imported, solved, and edited according to downstream design tasks. In addition, we condition the model on various contexts, including partial sketches (primers) and images of hand-drawn sketches. Evaluation of the proposed approach demonstrates its ability to synthesize realistic CAD sketches and its potential to aid the mechanical design workflow.

Observations of the Milky Way's low-α disk show that at fixed metallicity, [Fe/H], several element abundance, [X/Fe], correlate with age, with unique slopes and small scatters around the age-[X/Fe] relations. In this study, we turn to simulations to explore the age-[X/Fe] relations for the elements C, N, O, Mg, Si, S, and Ca that are traced in a FIRE-2 cosmological zoom-in simulation of a Milky Way-like galaxy, m12i, and understand what physical conditions give rise to the observed age-[X/Fe] trends. We first explore the distributions of mono-age populations in their birth and current locations, [Fe/H], and [X/Fe], and find evidence for inside-out radial growth for stars with ages < 7 Gyr. We then examine the age-[X/Fe] relations across m12i's disk and find that the direction of the trends agree with observations, apart from C, O, and Ca, with remarkably small intrinsic scatters, σint (0.01-0.04 dex). This σint measured in the simulations is also metallicity-dependent, with σint ≈ 0.025 dex at [Fe/H]=-0.25 dex versus σint ≈ 0.015 dex at [Fe/H]=0 dex, and a similar metallicity dependence is seen in the GALAH survey for the elements in common. Additionally, we find that σint is higher in the inner galaxy, where stars are older and formed in less chemically-homogeneous environments. The age-[X/Fe] relations and the small scatter around them indicate that simulations capture similar chemical enrichment variance as observed in the Milky Way, arising from stars sharing similar element abundances at a given birth place and time.

Uncertain feedback processes in galaxies affect the distribution of matter, currently limiting the power of weak lensing surveys. If we can identify cosmological statistics that are robust against these uncertainties, or constrain these effects by other means, then we can enhance the power of current and upcoming observations from weak lensing surveys such as DES, Euclid, the Rubin Observatory, and the Roman Space Telescope. In this work, we investigate the potential of the electron density auto-power spectrum as a robust probe of cosmology and baryonic feedback. We use a suite of (magneto-)hydrodynamic simulations from the CAMELS project and perform an idealized analysis to forecast statistical uncertainties on a limited set of cosmological and physically-motivated astrophysical parameters. We find that the electron number density auto-correlation, measurable through either kinematic Sunyaev-Zel'dovich observations or through Fast Radio Burst dispersion measures, provides tight constraints on Ωm and the mean baryon fraction in intermediate-mass halos, f¯bar. By obtaining an empirical measure for the associated systematic uncertainties, we find these constraints to be largely robust to differences in baryonic feedback models implemented in hydrodynamic simulations. We further discuss the main caveats associated with our analysis, and point out possible directions for future work.

Modern surveys of gravitational microlensing events have progressed to detecting thousands per year. Surveys are capable of probing Galactic structure, stellar evolution, lens populations, black hole physics, and the nature of dark matter. One of the key avenues for doing this is studying the microlensing Einstein radius crossing time distribution (tE). However, systematics in individual light curves as well as over-simplistic modeling can lead to biased results. To address this, we developed a model to simultaneously handle the microlensing parallax due to Earth's motion, systematic instrumental effects, and unlensed stellar variability with a Gaussian Process model. We used light curves for nearly 10,000 OGLE-III and IV Milky Way bulge microlensing events and fit each with our model. We also developed a forward model approach to infer the timescale distribution by forward modeling from the data rather than using point estimates from individual events. We find that modeling the variability in the baseline removes a source of significant bias in individual events, and previous analyses over-estimated the number of long timescale (tE>100 days) events due to their over simplistic models ignoring parallax effects and stellar variability. We use our fits to identify hundreds of events that are likely black holes.

Sub-millimeter emission lines produced by the interstellar medium (ISM) are strong tracers of star formation and are some of the main targets of line intensity mapping (LIM) surveys. In this work we present an empirical multi-line emission model that simultaneously covers the mean, scatter, and correlations of [CII], CO J=1-0 to J=5-4, and [CI] lines in redshift range 1≤z≤9. We assume the galaxy ISM line emission luminosity versus halo mass relations can be described by double power laws with redshift-dependent log normal scatter. The model parameters are then derived by fitting to the state of the art semi-analytic simulation results that have successfully reproduced multiple sub-millimeter line observations at 0≤z≲6. We cross check the line emission statistics predicted by the semi-analytic simulation and our empirical model, finding that at z≥1 our model reproduces the simulated line intensities with fractional error less than about 10%. The fractional difference is less than 25% for the power spectra. Grounded on physically-motivated and self-consistent galaxy simulations, this computationally efficient model will be helpful in forecasting ISM emission line statistics for upcoming LIM surveys.

Single-molecule force spectroscopy (smFS) is a powerful approach to studying molecular self-organization. However, the coupling of the molecule with the ever-present experimental device introduces artifacts, that complicate the interpretation of these experiments. Performing statistical inference to learn hidden molecular properties is challenging because these measurements produce non-Markovian time series, and even minimal models lead to intractable likelihoods. To overcome these challenges, we developed a computational framework built on novel statistical methods called simulation-based inference (SBI). SBI enabled us to directly estimate the Bayesian posterior, and extract reduced quantitative models from smFS, by encoding a mechanistic model into a simulator in combination with probabilistic deep learning. Using synthetic data, we could systematically disentangle the measurement of hidden molecular properties from experimental artifacts. The integration of physical models with machine-learning density estimation is general, transparent, easy to use, and broadly applicable to other types of biophysical experiments.

Single-molecule force spectroscopy (smFS) is a powerful approach to studying molecular self-organization. However, the coupling of the molecule with the ever-present experimental device introduces artifacts, that complicate the interpretation of these experiments. Performing statistical inference to learn hidden molecular properties is challenging because these measurements produce non-Markovian time series, and even minimal models lead to intractable likelihoods. To overcome these challenges, we developed a computational framework built on novel statistical methods called simulation-based inference (SBI). SBI enabled us to directly estimate the Bayesian posterior, and extract reduced quantitative models from smFS, by encoding a mechanistic model into a simulator in combination with probabilistic deep learning. Using synthetic data, we could systematically disentangle the measurement of hidden molecular properties from experimental artifacts. The integration of physical models with machine-learning density estimation is general, transparent, easy to use, and broadly applicable to other types of biophysical experiments.

The identification of the electromagnetic counterpart candidate ZTF19abanrhr to the binary black hole merger GW190521 opens the possibility to infer cosmological parameters from this standard siren with a uniquely identified host galaxy. The distant merger allows for cosmological inference beyond the Hubble constant. Here we show that the three-dimensional spatial location of ZTF19abanrhr calculated from the electromagnetic data remains consistent with the updated sky localization of GW190521 provided by the LIGO-Virgo Collaboration. If ZTF19abanrhr is associated with the GW190521 merger and assuming a flat wCDM model we find that H0=48+24−10 km/s/Mpc, Ωm=0.39+0.38−0.29, and w0=−1.29+0.63−0.50 (median and 68% credible interval). If we use the Hubble constant value inferred from another gravitational-wave event, GW170817, as a prior for our analysis, together with assumption of a flat ΛCDM and the model-independent constraint on the physical matter density ωm from Planck, we find H0=69.18.7−6.0 km/s/Mpc.

Extracting non-Gaussian information from the non-linear regime of structure formation is key to fully exploiting the rich data from upcoming cosmological surveys probing the large-scale structure of the universe. However, due to theoretical and computational complexities, this remains one of the main challenges in analyzing observational data. We present a set of summary statistics for cosmological matter fields based on 3D wavelets to tackle this challenge. These statistics are computed as the spatial average of the complex modulus of the 3D wavelet transform raised to a power q and are therefore known as invariant wavelet moments. The 3D wavelets are constructed to be radially band-limited and separable on a spherical polar grid and come in three types: isotropic, oriented, and harmonic. In the Fisher forecast framework, we evaluate the performance of these summary statistics on matter fields from the Quijote suite, where they are shown to reach state-of-the-art parameter constraints on the base ΛCDM parameters, as well as the sum of neutrino masses. We show that we can improve constraints by a factor 5 to 10 in all parameters with respect to the power spectrum baseline.

The total mass of the Local Group (LG) is a fundamental quantity that enables interpreting the orbits of its constituent galaxies and placing the LG in a cosmological context. One of the few methods that allows inferring the total mass directly is the "Timing Argument," which models the relative orbit of the Milky Way (MW) and M31. However, the MW itself is not in equilibrium, a byproduct of its merger history and the recent pericentric passage of the LMC/SMC. As a result, recent work has found that the MW disk is moving with a lower bound "travel velocity" of ∼32 km s−1 with respect to the outer stellar halo (Petersen & Peñarrubia 2021), thus biasing past Timing Argument measurements that do not account for this motion. We measure the total LG mass using a Timing Argument model that incorporates this measured travel velocity of the MW disk using several different compilations of recent kinematic measurements of M31. We find that incorporating the measured travel velocity lowers the inferred LG mass by 10-20 percent compared to a static MW halo, and find an updated total mass of either 4.0+0.5−0.3×1012M⊙ or 4.5+0.8−0.6×1012M⊙ depending on the adopted dataset. Measurements of the travel velocity with more distant tracers could yield even larger values, which would further decrease the inferred LG mass. Therefore, the newly measured travel velocity directly implies a lower LG mass than from a model with a static MW halo and must be considered in future dynamical studies of the Local Volume.