arXiv subjects

Thomas Müller

Publications and source records attributed to Thomas Müller.

At least 37 records · Page 2Linked to original sources

arXiv2022

Object classification on video data of meteors and meteor-like phenomena: algorithm and data

Every moment, countless meteoroids enter our atmosphere unseen. The detection and measurement of meteors offer the unique opportunity to gain insights into the composition of our solar systems' celestial bodies. Researchers, therefore, carry out a wide-area-sky-monitoring to secure 360-degree video material, saving every single entry of a meteor. Existing machine intelligence cannot accurately recognize events of meteors intersecting the earth's atmosphere due to a lack of high-quality training data publicly available. This work presents four reusable open source solutions for researchers trained on data we collected due to the lack of available labeled high-quality training data. We refer to the proposed dataset as the NightSkyUCP dataset, consisting of a balanced set of 10,000 meteor- and 10,000 non-meteor-events. Our solutions apply various machine learning techniques, namely classification, feature learning, anomaly detection, and extrapolation. For the classification task, a mean accuracy of 99.1\% is achieved. The code and data are made public at figshare with DOI: 10.6084/m9.figshare.16451625

astro-ph.IM↗

arXiv2022

Variable Bitrate Neural Fields

Neural approximations of scalar and vector fields, such as signed distance functions and radiance fields, have emerged as accurate, high-quality representations. State-of-the-art results are obtained by conditioning a neural approximation with a lookup from trainable feature grids that take on part of the learning task and allow for smaller, more efficient neural networks. Unfortunately, these feature grids usually come at the cost of significantly increased memory consumption compared to stand-alone neural network models. We present a dictionary method for compressing such feature grids, reducing their memory consumption by up to 100x and permitting a multiresolution representation which can be useful for out-of-core streaming. We formulate the dictionary optimization as a vector-quantized auto-decoder problem which lets us learn end-to-end discrete neural representations in a space where no direct supervision is available and with dynamic topology and structure. Our source code will be available at https://github.com/nv-tlabs/vqad.

cs.CV↗

arXiv2022

Predicted future fate of COSMOS galaxy protoclusters over 11 Gyr with constrained simulations

Cosmological simulations are crucial tools in studying the Universe, but they typically do not directly match real observed structures. Constrained cosmological simulations, on the other hand, are designed to match the observed distribution of galaxies. Here we present constrained simulations based on spectroscopic surveys at a redshift of z~2.3, corresponding to an epoch of nearly 11 Gyrs ago. This allows us to 'fast-forward' the simulation to our present-day and study the evolution of observed cosmic structures self-consistently. We confirm that several previously-reported protoclusters will evolve into massive galaxy clusters by our present epoch, including the 'Hyperion' structure that we predict will collapse into a giant filamentary supercluster spanning 100 Megaparsecs. We also discover previously unknown protoclusters, with lower final masses than typically detectable by other methods, that nearly double the number of known protoclusters within this volume. Constrained simulations, applied to future high-redshift datasets, represents a unique opportunity for studying early structure formation and matching galaxy properties between high and low redshifts.

astro-ph.CO↗

arXiv2022

Active Few-Shot Learning with FASL

Recent advances in natural language processing (NLP) have led to strong text classification models for many tasks. However, still often thousands of examples are needed to train models with good quality. This makes it challenging to quickly develop and deploy new models for real world problems and business needs. Few-shot learning and active learning are two lines of research, aimed at tackling this problem. In this work, we combine both lines into FASL, a platform that allows training text classification models using an iterative and fast process. We investigate which active learning methods work best in our few-shot setup. Additionally, we develop a model to predict when to stop annotating. This is relevant as in a few-shot setup we do not have access to a large validation set.

cs.CL↗

arXiv2022

Zero and Few-shot Learning for Author Profiling

Author profiling classifies author characteristics by analyzing how language is shared among people. In this work, we study that task from a low-resource viewpoint: using little or no training data. We explore different zero and few-shot models based on entailment and evaluate our systems on several profiling tasks in Spanish and English. In addition, we study the effect of both the entailment hypothesis and the size of the few-shot training sample. We find that entailment-based models out-perform supervised text classifiers based on roberta-XLM and that we can reach 80% of the accuracy of previous approaches using less than 50\% of the training data on average.

cs.CL↗

arXiv2022

Instant Neural Graphics Primitives with a Multiresolution Hash Encoding

Neural graphics primitives, parameterized by fully connected neural networks, can be costly to train and evaluate. We reduce this cost with a versatile new input encoding that permits the use of a smaller network without sacrificing quality, thus significantly reducing the number of floating point and memory access operations: a small neural network is augmented by a multiresolution hash table of trainable feature vectors whose values are optimized through stochastic gradient descent. The multiresolution structure allows the network to disambiguate hash collisions, making for a simple architecture that is trivial to parallelize on modern GPUs. We leverage this parallelism by implementing the whole system using fully-fused CUDA kernels with a focus on minimizing wasted bandwidth and compute operations. We achieve a combined speedup of several orders of magnitude, enabling training of high-quality neural graphics primitives in a matter of seconds, and rendering in tens of milliseconds at a resolution of ${1920\!\times\!1080}$.

cs.CV↗

arXiv2022

Few-Shot Learning with Siamese Networks and Label Tuning

We study the problem of building text classifiers with little or no training data, commonly known as zero and few-shot text classification. In recent years, an approach based on neural textual entailment models has been found to give strong results on a diverse range of tasks. In this work, we show that with proper pre-training, Siamese Networks that embed texts and labels offer a competitive alternative. These models allow for a large reduction in inference cost: constant in the number of labels rather than linear. Furthermore, we introduce label tuning, a simple and computationally efficient approach that allows to adapt the models in a few-shot setup by only changing the label embeddings. While giving lower performance than model fine-tuning, this approach has the architectural advantage that a single encoder can be shared by many different tasks.

cs.CL↗

arXiv2022

Stratospheric balloons as a platform for the next large far infrared observatory

Observations that require large physical instrument dimensions and/or a considerable amount of cryogens, as it is the case for high spatial resolution far infrared (FIR) astronomy, currently still face technological limits for their execution from space. Angular resolution and available observational capabilities are particularly affected. Balloon-based platforms promise to complement the existing observational capabilities by offering means to deploy comparatively large telescopes with comparatively little effort, including other advantages such as the possibility to regularly refill cryogens and to change and/or update instruments. The planned European Stratospheric Balloon Observatory (ESBO) aims at providing these additional large aperture FIR capabilities, exceeding the spatial resolution of Herschel, in the long term. The plans focus on reusable platforms performing regular flights and an operations concept that provides researchers with proposal-based access to observations. It thereby aims at offering a complement to other airborne, ground-based and space-based observatories in terms of access to wavelength regions, spatial resolution capability, and photometric stability. While the FIR capabilities are a main long-term objective, ESBO will offer benefits in other wavelength regimes along the way. Within the ESBO Design Study (ESBO DS), a prototype platform carrying a 0.5 m telescope for ultraviolet and visible light observations is being built and a platform concept for a next-generation FIR telescope is being studied. A flight of the UV/VIS prototype platform is estimated for 2021. In this paper we will outline the scientific and technical motivation for a large aperture balloon-based FIR observatory and the ESBO DS approach towards such an infrastructure. Secondly, we will present the technical motivation, science case, and instrumentation of the 0.5 m UV/VIS platform.

astro-ph.IM↗

arXiv2022

Stratospheric Balloons as a Complement to the Next Generation of Astronomy Missions

Observations that require large physical instrument dimensions and/or a considerable amount of cryogens, as it is for example the case for high spatial resolution far infrared astronomy, currently still face technological limits for their execution from space. The high cost and finality of space missions furthermore call for a very low risk approach and entail long development times. For certain spectral regions, prominently including the mid- to far-infrared as well as parts of the ultraviolet, stratospheric balloons offer a flexible and affordable complement to space telescopes, with short development times and comparatively good observing conditions. Yet, the entry burden to use balloon-borne telescopes is high, with research groups typically having to shoulder part of the infrastructure development as well. Aiming to ease access to balloon-based observations, we present the efforts towards a community-accessible balloon-based observatory, the European Stratospheric Balloon Observatory (ESBO). ESBO aims at complementing space-based and airborne capabilities over the next 10-15 years and at adding to the current landscape of scientific ballooning activities by providing a service-centered infrastructure for broader astronomical use, performing regular flights and offering an operations concept that provides researchers with a similar proposal-based access to observation time as practiced on ground-based observatories. We present details on the activities planned towards the goal of ESBO, the current status of the STUDIO (Stratospheric UV Demonstrator of an Imaging Observatory) prototype platform and mission, as well as selected technology developments with extensibility potential to space missions undertaken for STUDIO.

astro-ph.IM↗

arXiv2021

Path Guiding Using Spatio-Directional Mixture Models

We propose a learning-based method for light-path construction in path tracing algorithms, which iteratively optimizes and samples from what we refer to as spatio-directional Gaussian mixture models (SDMMs). In particular, we approximate incident radiance as an online-trained $5$D mixture that is accelerated by a $k$D-tree. Using the same framework, we approximate BSDFs as pre-trained $n$D mixtures, where $n$ is the number of BSDF parameters. Such an approach addresses two major challenges in path-guiding models. First, the $5$D radiance representation naturally captures correlation between the spatial and directional dimensions. Such correlations are present in e.g. parallax and caustics. Second, by using a tangent-space parameterization of Gaussians, our spatio-directional mixtures can perform approximate product sampling with arbitrarily oriented BSDFs. Existing models are only able to do this by either foregoing anisotropy of the mixture components or by representing the radiance field in local (normal aligned) coordinates, which both make the radiance field more difficult to learn. An additional benefit of the tangent-space parameterization is that each individual Gaussian is mapped to the solid sphere with low distortion near its center of mass. Our method performs especially well on scenes with small, localized luminaires that induce high spatio-directional correlation in the incident radiance.

cs.GR↗

arXiv2021

Towards a European Stratospheric Balloon Observatory -- The ESBO Design Study

This paper presents the concept of a community-accessible stratospheric balloon-based observatory that is currently under preparation by a consortium of European research institutes and industry. The planned European Stratospheric Balloon Observatory (ESBO) aims at complementing the current landscape of scientific ballooning activities by providing a service-centered infrastructure tailored towards broad astronomical use. In particular, the concept focuses on reusable platforms with exchangeable instruments and telescopes performing regular flights and an operations concept that provides researchers with options to test and operate own instruments, but later on also a proposal-based access to observations. It thereby aims at providing a complement to ground-, space-based, and airborne observatories in terms of access to wavelength regimes - particularly the ultraviolet (UV) and far infrared (FIR) regimes -, spatial resolution capability, and photometric stability. Within the currently ongoing ESBO Design Study (ESBO DS), financed within the European Union's Horizon 2020 Programme, a prototype platform carrying a 0.5-m telescope for UV and visible light observations is being built and concepts for larger following platforms, leading up to a next-generation FIR telescope are being studied. A flight of the UV/visible prototype platform is currently foreseen for 2021. We present the technical motivation, science case, instrumentation, and a two-stage image stabilization approach of the 0.5-m UV/visible platform. In addition, we briefly describe the novel mid-sized stabilized balloon gondola under design to carry telescopes in the 0.5 to 0.6 m range as well as the currently considered flight option for this platform. Secondly, we outline the scientific and technical motivation for a large balloon-based FIR telescope and the ESBO DS approach towards such an infrastructure.

astro-ph.IM↗

arXiv2021

MATE: Multi-view Attention for Table Transformer Efficiency

This work presents a sparse-attention Transformer architecture for modeling documents that contain large tables. Tables are ubiquitous on the web, and are rich in information. However, more than 20% of relational tables on the web have 20 or more rows (Cafarella et al., 2008), and these large tables present a challenge for current Transformer models, which are typically limited to 512 tokens. Here we propose MATE, a novel Transformer architecture designed to model the structure of web tables. MATE uses sparse attention in a way that allows heads to efficiently attend to either rows or columns in a table. This architecture scales linearly with respect to speed and memory, and can handle documents containing more than 8000 tokens with current accelerators. MATE also has a more appropriate inductive bias for tabular data, and sets a new state-of-the-art for three table reasoning datasets. For HybridQA (Chen et al., 2020b), a dataset that involves large documents containing tables, we improve the best prior result by 19 points.

cs.CL↗

arXiv2021

Dynamic Diffuse Global Illumination Resampling

Interactive global illumination remains a challenge in radiometrically- and geometrically-complex scenes. Specialized sampling strategies are effective for specular and near-specular transport because the scattering has relatively low directional variance per scattering event. In contrast, the high variance from transport paths comprising multiple rough glossy or diffuse scattering events remains notoriously difficult to resolve with a small number of samples. We extend unidirectional path tracing to address this by combining screen-space reservoir resampling and sparse world-space probes, significantly improving sample efficiency for transport contributions that terminate on diffuse scattering events. Our experiments demonstrate a clear improvement -- at equal time and equal quality -- over purely path traced and purely probe-based baselines. Moreover, when combined with commodity denoisers, we are able to interactively render global illumination in complex scenes.

cs.GR↗

arXiv2021

Real-time Neural Radiance Caching for Path Tracing

We present a real-time neural radiance caching method for path-traced global illumination. Our system is designed to handle fully dynamic scenes, and makes no assumptions about the lighting, geometry, and materials. The data-driven nature of our approach sidesteps many difficulties of caching algorithms, such as locating, interpolating, and updating cache points. Since pretraining neural networks to handle novel, dynamic scenes is a formidable generalization challenge, we do away with pretraining and instead achieve generalization via adaptation, i.e. we opt for training the radiance cache while rendering. We employ self-training to provide low-noise training targets and simulate infinite-bounce transport by merely iterating few-bounce training updates. The updates and cache queries incur a mild overhead -- about 2.6ms on full HD resolution -- thanks to a streaming implementation of the neural network that fully exploits modern hardware. We demonstrate significant noise reduction at the cost of little induced bias, and report state-of-the-art, real-time performance on a number of challenging scenarios.

cs.GR↗

arXiv2021

Open Domain Question Answering over Tables via Dense Retrieval

Recent advances in open-domain QA have led to strong models based on dense retrieval, but only focused on retrieving textual passages. In this work, we tackle open-domain QA over tables for the first time, and show that retrieval can be improved by a retriever designed to handle tabular context. We present an effective pre-training procedure for our retriever and improve retrieval quality with mined hard negatives. As relevant datasets are missing, we extract a subset of Natural Questions (Kwiatkowski et al., 2019) into a Table QA dataset. We find that our retriever improves retrieval results from 72.0 to 81.1 recall@10 and end-to-end QA results from 33.8 to 37.7 exact match, over a BERT based retriever.

cs.CL↗

arXiv2021

DoT: An efficient Double Transformer for NLP tasks with tables

Transformer-based approaches have been successfully used to obtain state-of-the-art accuracy on natural language processing (NLP) tasks with semi-structured tables. These model architectures are typically deep, resulting in slow training and inference, especially for long inputs. To improve efficiency while maintaining a high accuracy, we propose a new architecture, DoT, a double transformer model, that decomposes the problem into two sub-tasks: A shallow pruning transformer that selects the top-K tokens, followed by a deep task-specific transformer that takes as input those K tokens. Additionally, we modify the task-specific attention to incorporate the pruning scores. The two transformers are jointly trained by optimizing the task-specific loss. We run experiments on three benchmarks, including entailment and question-answering. We show that for a small drop of accuracy, DoT improves training and inference time by at least 50%. We also show that the pruning transformer effectively selects relevant tokens enabling the end-to-end model to maintain similar accuracy as slower baseline models. Finally, we analyse the pruning and give some insight into its impact on the task model.

cs.CL↗

arXiv2021

Measurement-induced dark state phase transitions in long-ranged fermion systems

We identify an unconventional algebraic scaling phase in the quantum dynamics of free fermions with long range hopping, which are exposed to continuous local density measurements. The unconventional phase is characterized by an algebraic entanglement entropy growth, and by a slow algebraic decay of the density-density correlation function, both with a fractional exponent. It occurs for hopping decay exponents $1< p \lesssim 3/2$ independently of the measurement rate. The algebraic phase gives rise to two critical lines, separating it from a critical phase with logarithmic entanglement growth at small, and an area law phase with constant entanglement entropy at large monitoring rates. A perturbative renormalization group analysis suggests that the transitions to the long-range phase are also unconventional, corresponding to a modified sine-Gordon theory. Comparing exact numerical simulations of the monitored wave functions with analytical predictions from a replica field theory approach yields an excellent quantitative agreement. This confirms the view of a measurement-induced phase transition as a quantum phase transition in the dark state of an effective, non-Hermitian Hamiltonian.

cond-mat.stat-mech↗

arXiv2021

Shape effects of localized losses in quantum wires: dissipative resonances and nonequilibrium universality

We study the effects of the spacial structure of localized losses in weakly-interacting fermionic quantum wires. We show that multiple dissipative impurities give rise to resonant effects visible in the transport properties and the particles' momentum distribution. These resonances can enhance or suppress the effective particle losses in the wire. Moreover, we investigate the interplay between interactions and the impurity shape and find that, differently from the coherent scatterer case, the impurity shape modifies the scaling of the scattering probabilities close to the Fermi momentum. We show that, while the fluctuation-induced quantum Zeno effect is robust against the shape of the impurities, the fluctuation-induced transparency is lifted continuously. This is reflected in the emergence of a continuous line of fixed points in the renormalization group flow of the scattering probabilities.

cond-mat.quant-gas↗