The local dark matter density determines the strength of the signal expected in direct-detection experiments, yet published estimates from stellar motions disagree by more than their errors, and the most recent machine-learning analysis of Gaia data finds a local density consistent with zero. According to Jeans' theorem, a distribution function built from integrals of motion satisfies the collisionless Boltzmann equation (CBE) trivially for any choice of potential, so a search that simultaneously fits the distribution function and the potential to the CBE identifies neither. Our pipeline instead estimates the distribution function in isolation, linearizing the equation in terms of accelerations and allowing for direct measurement of the local force field, and then fits closed forms to that field via symbolic regression. Throughout, we find that the usable information lies not in the CBE residual but in the stellar number counts, the observable most distorted by survey selection. Along the vertical profile, our recovered potential agrees with the classical self-gravitating isothermal disc.
In many scientific disciplines, weak signals of interest are obscured by dominant nuisance signals that are several orders of magnitude stronger. Recovering these weak signals requires subtracting the dominant ones; however, this calibration process inherently distorts or partially suppresses the underlying signal of interest. To address this problem, we propose the use of normalizing flow models to reconstruct calibration-affected weak signals. By leveraging the statistical invariance of the target signals and assuming minimal initial suppression, our framework effectively recovers the lost signal components. We provide a comprehensive theoretical overview of this normalizing flow-based recovery method and demonstrate its efficacy using simulated data.
Astronomical spectra encode rich physical information, but drawing scientific conclusions from spectral features typically requires expert interpretation. This paper presents AstroSpecLM, a spectrum-language model that connects one-dimensional DESI spectra with Qwen3-4B to answer questions and provide explanations grounded in spectral evidence. Instead of generating question-answer pairs directly from templates or raw catalog fields, we first distill each spectrum into a compact set of catalog- and spectrum-derived facts, then use these facts as references to generate instruction-following conversations. The resulting model is competitive with specialist supervised baselines on classification and redshift estimation, while additionally producing natural-language explanations that reference specific spectral features. Our results indicate that grounding a language model in one-dimensional scientific spectra is feasible, and that fact-mediated instruction data yields a model capable of both prediction and explanation.
Large-volume neutrino telescopes infer neutrino properties from Cherenkov light, but simulating the transport of billions of photons through highly scattering ice or water is computationally costly. We introduce candela, a differentiable SIREN neural field that learns the photon Green's function of the IceCube Neutrino Observatory, a cubic-kilometer detector embedded in Antarctic glacial ice. Given a point-like energy deposit and sensor, it predicts the expected photon yield and full arrival-time distribution at the sensor. Complete events are simulated by decomposing charged-particle energy deposits into point-like sources and superposing their predicted sensor responses. Trained on Monte-Carlo simulations, candela generates events $50$--$100\times$ faster than existing methods, with cost scaling only weakly with neutrino energy. It keeps median yields within $2\%$ of the MC expectation and timing distributions at the MC statistical floor across six photon-count decades. The model also provides end-to-end gradients with respect to event parameters and opens a path toward optimizing scattering-medium properties, which often dominate systematic uncertainties in neutrino telescopes.
We present a deconvolution-based framework for testing scale-dependent departures of the late-time matter power spectrum from a fiducial cosmological model using $3\times2$pt measurements. We introduce a free-form scale-dependent modulation $A(k)$ of the fiducial nonlinear matter power spectrum and construct the linear response of binned galaxy-clustering, galaxy-galaxy-lensing, and cosmic-shear spectra to the discretized modulation $A(k)$. The response is evaluated with full-sky, beyond-Limber kernels including density, redshift-space-distortion, gravitational-shear, and intrinsic-alignment contributions. We reconstruct $A(k)$ using a regularized modified Richardson-Lucy algorithm, with weak diffusion in $\ln k$ and selection of the minimum-$χ^2$ solution along the iteration history. Using Rubin/LSST Year 10-like synthetic data, we find that oscillatory modulations with amplitudes $\gtrsim1\%$ can be recovered over $0.1\lesssim k\lesssim0.5\,{\rm Mpc}^{-1}$, provided the oscillation frequency $f\lesssim10$ on $\log_{10}[k/(0.2\,{\rm Mpc}^{-1})]$. We further introduce a posterior-weighted consistency statistic calibrated with posterior-predictive null mocks, thereby accounting for cosmological and nuisance-parameter uncertainties without relying on Wilks' theorem. The null case is consistent with $A(k)=1$, while a $1\%$ oscillatory modulation is detected at $\sim 2.6σ$. These results demonstrate the potential of regularized deconvolution as a model-independent consistency test of the matter power spectrum in future $3\times2$pt surveys.
The matter power spectrum is one of the primary statistical descriptors of the large-scale distribution of matter in the Universe and provides a powerful probe of cosmic structure formation. Measurements of cosmic microwave background (CMB) lensing offer an integrated view of the matter distribution over a wide range of redshifts, enabling the reconstruction of the underlying matter power spectrum. In this work, we reconstruct the reference linear matter power spectrum $ P_\text{lin}(k,0)$ from the baseline joint CMB lensing measurements of Planck PR4, ACT DR6, and SPT-3G using a covariance-weighted modified Richardson-Lucy(MRL) deconvolution algorithm. The reconstructed spectrum is found to be consistent with the fiducial linear prediction on large scales, while exhibiting a systematic enhancement for $k \gtrsim 0.1\,{\rm Mpc}^{-1}$, where nonlinear gravitational evolution becomes important. To investigate this behavior, we introduce a scale-dependent correction factor, $A(k)$, defined through $P(k)=A(k)\,P_{\rm nl}(k),$ where $P_{\rm nl}(k)$ is the fiducial nonlinear matter power spectrum obtained from 2LPT simulations. The reconstructed correction factor remains consistent with unity within $2σ$ confidence over the reconstructed range, indicating that the observed enhancement is well explained by the standard nonlinear evolution of the matter power spectrum. In addition, the reconstruction shows agreement with the fiducial BAO template around the BAO feature at $k\sim(0.04-0.06)\ {\rm Mpc}^{-1}$, indicating that some BAO-scale information survives the lensing projection.
Aims. The objective of this work is to adapt a deep neural network model to perform galaxy morphological classification trained from crowd annotations, considering the training scheme, the agreement between the annotators, and the hierarchy. Methods. We use Galaxy Zoo 1 as our experimental testbed and trained a convolutional neural network (CNN) for the automatic classification of galaxies' morphologies. We analyze the impact of the following aspects on the classification accuracy and training efficiency: (i) Training only the last layer vs. training all the network; (ii) Classification with only the CNN vs. considering the hierarchy; (iii) Comparing the models trained with different amounts of data and levels of agreement between the annotators; (iv) Training by stages, transferring knowledge from one model to another; and (v) Combining several models as an ensemble. Results From the experiments, we derive the following results: (i) Training all the layers in the network significantly improves the accuracy (10% increase in exact match), compared to training only the last layer; (ii) There is a tradeoff between the amount of data and the level of agreement between the annotators used for training; (iii) Using the hierarchy can improve accuracy when the amount of training data is reduced; (iv) Training by stages through transfer learning (curriculum learning) produces higher accuracy for limited data; (v) Ensembles can improve accuracy; (vi) Models achieve a low accuracy for the most difficult cases, but, if we consider hierarchical measures, we can derive useful results for upper levels in the hierarchy. An accuracy above 99% is achieved when training all layers of the network and considering a high agreement between the annotators. Conclusions. Training deep learning models from crowd annotations involves additional challenges than learning from hard annotations.