arXiv subjects

Daan Roos

Publications and source records attributed to Daan Roos.

3 recordsLinked to original sources

arXiv2026

Categorical Flow Maps

We introduce Categorical Flow Maps, a flow-matching method for accelerated few-step generation of categorical data via self-distillation. Building on recent variational formulations of flow matching and the broader trend towards accelerated inference in diffusion and flow-based models, we define a flow map towards the simplex that transports probability mass toward a predicted endpoint, yielding a parametrisation that naturally constrains model predictions. Since our trajectories are continuous rather than discrete, Categorical Flow Maps can be trained with existing distillation techniques, as well as a new objective based on endpoint consistency. This continuous formulation also automatically unlocks test-time inference: we can directly reuse existing guidance and reweighting techniques in the categorical setting to steer sampling toward downstream objectives. Empirically, we achieve state-of-the-art few-step results on images, molecular graphs, and text, with strong performance even in single-step generation.

cs.LG↗

arXiv2025

Lens based Kinetic Inductance Detectors with Distributed Dual Polarised Absorbers for Far Infra-red Space-based Astronomy

Future space-based far infra-red astronomical observations require background limited detector sensitivities and scalable focal plane array solutions to realise their vast potential in observation speed. In this work, a focal plane array of lens absorber coupled Kinetic Inductance Detectors (KIDs) is proposed to fill this role. The figures of merit and design guidelines for the proposed detector concept are derived by employing a previously developed electromagnetic spectral modelling technique. Two designs operating at central frequencies of $6.98$ and $12$ THz are studied. A prototype array of the former is fabricated, and its performance is experimentally determined and validated. Specifically, the optical coupling of the detectors to incoherent distributed sources (i.e. normalised throughput) is quantified experimentally with good agreement with the estimations provided by the model. The coupling of the lens absorber prototypes to an incident plane wave, i.e. aperture efficiency, is also indirectly validated experimentally matching the expected value of $54\%$ averaged over two polarisation. The noise equivalent power of the KIDs are also measured with limiting value of $8\times10^{-20}$ $\mathrm{W/\sqrt{Hz}}$.

astro-ph.IM↗

arXiv2022

Deterministic training of generative autoencoders using invertible layers

In this work, we provide a deterministic alternative to the stochastic variational training of generative autoencoders. We refer to these new generative autoencoders as AutoEncoders within Flows (AEF), since the encoder and decoder are defined as affine layers of an overall invertible architecture. This results in a deterministic encoding of the data, as opposed to the stochastic encoding of VAEs. The paper introduces two related families of AEFs. The first family relies on a partition of the ambient space and is trained by exact maximum-likelihood. The second family exploits a deterministic expansion of the ambient space and is trained by maximizing the log-probability in this extended space. This latter case leaves complete freedom in the choice of encoder, decoder and prior architectures, making it a drop-in replacement for the training of existing VAEs and VAE-style models. We show that these AEFs can have strikingly higher performance than architecturally identical VAEs in terms of log-likelihood and sample quality, especially for low dimensional latent spaces. Importantly, we show that AEF samples are substantially sharper than VAE samples.

stat.ML↗