We proposed an extension of Akaike's relative power contribution that could be applied to data with correlations between noises. This method decomposes the power spectrum into a contribution of the terms caused by correlation between two noises, in addition to the contributions of the independent noises. Numerical examples confirm that some of the correlated noise has the effect of reducing the power spectrum.
相關內容
Datalogo is an extension of Datalog that allows for aggregation and recursion over an arbitrary commutative semiring. Like Datalog, Datalogo programs can be evaluated via the natural iterative algorithm until a fixed point is reached. However unlike Datalog, the natural iterative evaluation of some Datalogo programs over some semirings may not converge. It is known that the commutative semirings for which the iterative evaluation of Datalogo programs is guaranteed to converge are exactly those semirings that are stable~\cite{Khamis0PSW22}. Previously, the best known upper bound on the number of iterations until convergence over $p$-stable semirings is $\sum_{i=1}^n (p+2)^i = \Theta(p^n)$ steps, where $n$ is (essentially) the output size. We establish that, in fact, the natural iterative evaluation of a Datalogoprogram over a $p$-stable semiring converges within a polynomial number of iterations. In particular our upper bound is $O( \sigma p n^2( n^2 \lg \lambda + \lg \sigma))$ where $\sigma$ is the number of elements in the semiring present in either the input databases or the Datalogo program, and $\lambda$ is the maximum number of terms in any product in the Datalogo program.
The modifiable areal unit problem in geography or the change-of-support (COS) problem in statistics demonstrates that the interpretation of spatial (or spatio-temporal) data analysis is affected by the choice of resolutions or geographical units used in the study. The ecological fallacy is one famous example of this phenomenon. Here we investigate the ecological fallacy associated with the COS problem for multivariate spatial data with the goal of providing a data-driven discretization criterion for the domain of interest that minimizes aggregation errors. The discretization is based on a novel multiscale metric, called the Multivariate Criterion for Aggregation Error (MVCAGE). Such multi-scale representations of an underlying multivariate process are often formulated in terms of basis expansions. We show that a particularly useful basis expansion in this context is the multivariate Karhunen-Lo`eve expansion (MKLE). We use the MKLE to build the MVCAGE loss function and use it within the framework of spatial clustering algorithms to perform optimal spatial aggregation. We demonstrate the effectiveness of our approach through simulation and through regionalization of county-level income and hospital quality data over the United States and prediction of ocean color in the coastal Gulf of Alaska.
Recent advances in natural language processing (NLP) owe their success to pre-training language models on large amounts of unstructured data. Still, there is an increasing effort to combine the unstructured nature of LMs with structured knowledge and reasoning. Particularly in the rapidly evolving field of biomedical NLP, knowledge-enhanced language models (KELMs) have emerged as promising tools to bridge the gap between large language models and domain-specific knowledge, considering the available biomedical knowledge graphs (KGs) curated by experts over the decades. In this paper, we develop an approach that uses lightweight adapter modules to inject structured biomedical knowledge into pre-trained language models (PLMs). We use two large KGs, the biomedical knowledge system UMLS and the novel biochemical ontology OntoChem, with two prominent biomedical PLMs, PubMedBERT and BioLinkBERT. The approach includes partitioning knowledge graphs into smaller subgraphs, fine-tuning adapter modules for each subgraph, and combining the knowledge in a fusion layer. We test the performance on three downstream tasks: document classification,question answering, and natural language inference. We show that our methodology leads to performance improvements in several instances while keeping requirements in computing power low. Finally, we provide a detailed interpretation of the results and report valuable insights for future work.
We study the problem of estimating the number of defective items in adaptive Group testing by using a minimum number of queries. We improve the existing algorithm and prove a lower bound that show that, for constant estimation, the number of tests in our algorithm is optimal.
The increasing digitization of smart grids has made addressing cybersecurity issues crucial in order to secure the power supply. Anomaly detection has emerged as a key technology for cybersecurity in smart grids, enabling the detection of unknown threats. Many research efforts have proposed various machine-learning-based approaches for anomaly detection in grid operations. However, there is a need for a reproducible and comprehensive evaluation environment to investigate and compare different approaches to anomaly detection. The assessment process is highly dependent on the specific application and requires an evaluation that considers representative datasets from the use case as well as the specific characteristics of the use case. In this work, we present an evaluation environment for anomaly detection methods in smart grids that facilitates reproducible and comprehensive evaluation of different anomaly detection methods.
Welch's method provides an estimator of the power spectral density that is statistically consistent. This is achieved by averaging over periodograms calculated from overlapping segments of a time series. For a finite length time series, while the variance of the estimator decreases as the number of segments increase, the magnitude of the estimator's bias increases: a bias-variance trade-off ensues when setting the segment number. We address this issue by providing a a novel method for debiasing Welch's method which maintains the computational complexity and asymptotic consistency, and leads to improved finite-sample performance. Theoretical results are given for fourth-order stationary processes with finite fourth-order moments and absolutely continuous fourth-order cumulant spectrum. The significant bias reduction is demonstrated with numerical simulation and an application to real-world data, where several empirical metrics indicate our debiased estimator compares favourably to Welch's. Our estimator also permits irregular spacing over frequency and we demonstrate how this may be employed for signal compression and further variance reduction. Code accompanying this work is available in the R and python languages.
Reg-ROMs are stabilization strategies that leverage spatial filtering to alleviate the spurious numerical oscillations generally displayed by the classical G-ROM in under-resolved numerical simulations of turbulent flows. In this paper, we propose a new Reg-ROM, the time-relaxation ROM (TR-ROM), which filters the marginally resolved scales. We compare the new TR-ROM with the two other Reg-ROMs in current use, i.e., the L-ROM and the EFR-ROM, in the numerical simulation of the turbulent channel flow at $Re_{\tau} = 180$ and $Re_{\tau} = 395$ in both the reproduction and the predictive regimes. For each Reg-ROM, we investigate two different filters: (i) the differential filter (DF), and (ii) a new higher-order algebraic filter (HOAF). In our numerical investigation, we monitor the Reg-ROM performance for the ROM dimension, $N$, and the filter order. We also perform sensitivity studies of the three Reg-ROMs for the time interval, relaxation parameter, and filter radius. The numerical results yield the following conclusions: (i) All three Reg-ROMs are significantly more accurate than the G-ROM and (ii) more accurate than the ROM projection, representing the best theoretical approximation of the training data in the given ROM space. (iii) With the optimal parameter values, the TR-ROM is more accurate than the other two Reg-ROMs in all tests. (iv) For most $N$ values, DF yields the most accurate results for all three Reg-ROMs. (v) The optimal parameters trained in the reproduction regime are also optimal for the predictive regime for most $N$ values. (vi) All three Reg-ROMs are sensitive to the filter radius and the filter order, and the EFR-ROM and the TR-ROM are sensitive to the relaxation parameter. (vii) The optimal range for the filter radius and the effect of relaxation parameter are similar for the two $\rm Re_\tau$ values.
This article provides quasi-optimal a priori error estimates for an optimal control problem constrained by an elliptic obstacle problem where the finite element discretization is carried out using the symmetric interior penalty discontinuous Galerkin method. The main proofs are based on the improved $L^2$-error estimates for the obstacle problem, the discrete maximum principle, and a well-known quadratic growth property. The standard (restrictive) assumptions on mesh are not assumed here.
The dominating NLP paradigm of training a strong neural predictor to perform one task on a specific dataset has led to state-of-the-art performance in a variety of applications (eg. sentiment classification, span-prediction based question answering or machine translation). However, it builds upon the assumption that the data distribution is stationary, ie. that the data is sampled from a fixed distribution both at training and test time. This way of training is inconsistent with how we as humans are able to learn from and operate within a constantly changing stream of information. Moreover, it is ill-adapted to real-world use cases where the data distribution is expected to shift over the course of a model's lifetime. The first goal of this thesis is to characterize the different forms this shift can take in the context of natural language processing, and propose benchmarks and evaluation metrics to measure its effect on current deep learning architectures. We then proceed to take steps to mitigate the effect of distributional shift on NLP models. To this end, we develop methods based on parametric reformulations of the distributionally robust optimization framework. Empirically, we demonstrate that these approaches yield more robust models as demonstrated on a selection of realistic problems. In the third and final part of this thesis, we explore ways of efficiently adapting existing models to new domains or tasks. Our contribution to this topic takes inspiration from information geometry to derive a new gradient update rule which alleviate catastrophic forgetting issues during adaptation.
With the rapid increase of large-scale, real-world datasets, it becomes critical to address the problem of long-tailed data distribution (i.e., a few classes account for most of the data, while most classes are under-represented). Existing solutions typically adopt class re-balancing strategies such as re-sampling and re-weighting based on the number of observations for each class. In this work, we argue that as the number of samples increases, the additional benefit of a newly added data point will diminish. We introduce a novel theoretical framework to measure data overlap by associating with each sample a small neighboring region rather than a single point. The effective number of samples is defined as the volume of samples and can be calculated by a simple formula $(1-\beta^{n})/(1-\beta)$, where $n$ is the number of samples and $\beta \in [0,1)$ is a hyperparameter. We design a re-weighting scheme that uses the effective number of samples for each class to re-balance the loss, thereby yielding a class-balanced loss. Comprehensive experiments are conducted on artificially induced long-tailed CIFAR datasets and large-scale datasets including ImageNet and iNaturalist. Our results show that when trained with the proposed class-balanced loss, the network is able to achieve significant performance gains on long-tailed datasets.
小貼士
登錄享
相關主題
注冊地址： 北京市海淀區羊坊店路18號2幢3層301-191