Estimating the structure of directed acyclic graphs (DAGs) of features (variables) plays a vital role in revealing the latent data generation process and providing causal insights in various applications. Although there have been many studies on structure learning with various types of data, the structure learning on the dynamic graph has not been explored yet, and thus we study the learning problem of node feature generation mechanism on such ubiquitous dynamic graph data. In a dynamic graph, we propose to simultaneously estimate contemporaneous relationships and time-lagged interaction relationships between the node features. These two kinds of relationships form a DAG, which could effectively characterize the feature generation process in a concise way. To learn such a DAG, we cast the learning problem as a continuous score-based optimization problem, which consists of a differentiable score function to measure the validity of the learned DAGs and a smooth acyclicity constraint to ensure the acyclicity of the learned DAGs. These two components are translated into an unconstraint augmented Lagrangian objective which could be minimized by mature continuous optimization techniques. The resulting algorithm, named GraphNOTEARS, outperforms baselines on simulated data across a wide range of settings that may encounter in real-world applications. We also apply the proposed approach on two dynamic graphs constructed from the real-world Yelp dataset, demonstrating our method could learn the connections between node features, which conforms with the domain knowledge.
相關內容
Gradient methods are experiencing a growth in methodological and theoretical developments owing to the challenges posed by optimization problems arising in data science. However, such gradient methods face diverging optimality gaps or exploding objective evaluations when applied to optimization problems with realistic properties for data science applications. In this work, we address this gap by developing a generic methodology that economically uses objective function evaluations in a problem-driven manner to prevent optimality gap divergence and avoid explosions in objective evaluations. Our methodology allows for a variety of step size routines and search direction strategies. Furthermore, we develop a particular, novel step size selection methodology that is well-suited to our framework. We show that our specific procedure is highly competitive with standard optimization methods on CUTEst test problems. We then show our specific procedure is highly favorable relative to standard optimization methods on a particularly tough data science problem: learning the parameters in a generalized estimating equation model. Thus, we provide a novel gradient methodology that is better suited to optimization problems from this important class of data science applications.
Current legal outcome prediction models - a staple of legal NLP - do not explain their reasoning. However, to employ these models in the real world, human legal actors need to be able to understand the model's decisions. In the case of common law, legal practitioners reason towards the outcome of a case by referring to past case law, known as precedent. We contend that precedent is, therefore, a natural way of facilitating explainability for legal NLP models. In this paper, we contribute a novel method for identifying the precedent employed by legal outcome prediction models. Furthermore, by developing a taxonomy of legal precedent, we are able to compare human judges and neural models with respect to the different types of precedent they rely on. We find that while the models learn to predict outcomes reasonably well, their use of precedent is unlike that of human judges.
Spiking Neural Networks (SNNs) and neuromorphic models are more efficient and have more biological realism than the activation functions typically used in deep neural networks, transformer models and generative AI. SNNs have local learning rules, are able to learn on small data sets, and can adapt through neuromodulation. Although research has shown their advantages, there are still few compelling practical applications, especially at the edge where sensors and actuators need to be processed in a timely fashion. One reason for this might be that SNNs are much more challenging to understand, build, and operate due to their intrinsic properties. For instance, the mathematical foundation involves differential equations rather than basic activation functions. To address these challenges, we have developed CARLsim++. It is an integrated toolbox that enables fast and easy creation of neuromorphic applications. It encapsulates the mathematical intrinsics and low-level C++ programming by providing a graphical user interface for users who do not have a background in software engineering but still want to create neuromorphic models. Developers can easily configure inputs and outputs to devices and robots. These can be accurately simulated before deploying on physical devices. CARLsim++ can lead to rapid development of neuromorphic applications for simulation or edge processing.
We extend the theory of locally checkable labeling problems (LCLs) from the classical LOCAL model to a number of other models that have been studied recently, including the quantum-LOCAL model, finitely-dependent processes, non-signaling model, dynamic-LOCAL model, and online-LOCAL model [e.g. STOC 2024, ICALP 2023]. First, we demonstrate the advantage that finitely-dependent processes have over the classical LOCAL model. We show that all LCL problems solvable with locality $O(\log^\star n)$ in the LOCAL model admit a finitely-dependent distribution (with constant locality). In particular, this gives a finitely-dependent coloring for regular trees, answering an open question by Holroyd [2023]. This also introduces a new formal barrier for understanding the distributed quantum advantage: it is not possible to exclude quantum advantage for any LCL in the $\Theta(\log^\star n)$ complexity class by using non-signaling arguments. Second, we put limits on the capabilities of all of these models. To this end, we introduce a model called randomized online-LOCAL, which is strong enough to simulate e.g. SLOCAL and dynamic-LOCAL, and we show that it is also strong enough to simulate any non-signaling distribution and hence any quantum-LOCAL algorithm. We prove the following result for rooted trees: if we can solve an LCL problem with locality $o(\log \log n)$ in the randomized online-LOCAL model, we can solve it with locality $O(\log^\star n)$ in the classical deterministic LOCAL model. Put together, these results show that in rooted trees the set of LCLs that can be solved with locality $O(\log^\star n)$ is the same across all these models: classical deterministic and randomized LOCAL, quantum-LOCAL, non-signaling model, dynamic-LOCAL, and deterministic and randomized online-LOCAL.
Object detection is critical in autonomous driving, and it is more practical yet challenging to localize objects of unknown categories: an endeavour known as Class-Agnostic Object Detection (CAOD). Existing studies on CAOD predominantly rely on ordinary cameras, but these frame-based sensors usually have high latency and limited dynamic range, leading to safety risks in real-world scenarios. In this study, we turn to a new modality enabled by the so-called event camera, featured by its sub-millisecond latency and high dynamic range, for robust CAOD. We propose Detecting Every Object in Events (DEOE), an approach tailored for achieving high-speed, class-agnostic open-world object detection in event-based vision. Built upon the fast event-based backbone: recurrent vision transformer, we jointly consider the spatial and temporal consistencies to identify potential objects. The discovered potential objects are assimilated as soft positive samples to avoid being suppressed as background. Moreover, we introduce a disentangled objectness head to separate the foreground-background classification and novel object discovery tasks, enhancing the model's generalization in localizing novel objects while maintaining a strong ability to filter out the background. Extensive experiments confirm the superiority of our proposed DEOE in comparison with three strong baseline methods that integrate the state-of-the-art event-based object detector with advancements in RGB-based CAOD. Our code is available at //github.com/Hatins/DEOE.
Aspect Category Detection (ACD) aims to identify implicit and explicit aspects in a given review sentence. The state-of-the-art approaches for ACD use Deep Neural Networks (DNNs) to address the problem as a multi-label classification task. However, learning category-specific representations heavily rely on the amount of labeled examples, which may not readily available in real-world scenarios. In this paper, we propose a novel approach to tackle the ACD task by combining DNNs with Gradual Machine Learning (GML) in a supervised setting. we aim to leverage the strength of DNN in semantic relation modeling, which can facilitate effective knowledge transfer between labeled and unlabeled instances during the gradual inference of GML. To achieve this, we first analyze the learned latent space of the DNN to model the relations, i.e., similar or opposite, between instances. We then represent these relations as binary features in a factor graph to efficiently convey knowledge. Finally, we conduct a comparative study of our proposed solution on real benchmark datasets and demonstrate that the GML approach, in collaboration with DNNs for feature extraction, consistently outperforms pure DNN solutions.
The Church Problem asks for the construction of a procedure which, given a logical specification A(I,O) between input omega-strings I and output omega-strings O, determines whether there exists an operator F that implements the specification in the sense that A(I, F(I)) holds for all inputs I. Buchi and Landweber provided a procedure to solve the Church problem for MSO specifications and operators computable by finite-state automata. We investigate a generalization of the Church synthesis problem to the continuous time of the non-negative reals. We show that in the continuous time there are phenomena which are very different from the canonical discrete time domain of the natural numbers.
Image forgery is a topic that has been studied for many years. Before the breakthrough of deep learning, forged images were detected using handcrafted features that did not require training. These traditional methods failed to perform satisfactorily even on datasets much worse in quality than real-life image manipulations. Advances in deep learning have impacted image forgery detection as much as they have impacted other areas of computer vision and have improved the state of the art. Deep learning models require large amounts of labeled data for training. In the case of image forgery, labeled data at the pixel level is a very important factor for the models to learn. None of the existing datasets have sufficient size, realism and pixel-level labeling at the same time. This is due to the high cost of producing and labeling quality images. It can take hours for an image editing expert to manipulate just one image. To bridge this gap, we automate data generation using image composition techniques that are very related to image forgery. Unlike other automated data generation frameworks, we use state of the art image composition deep learning models to generate spliced images close to the quality of real-life manipulations. Finally, we test the generated dataset on the SOTA image manipulation detection model and show that its prediction performance is lower compared to existing datasets, i.e. we produce realistic images that are more difficult to detect. Dataset will be available at //github.com/99eren99/DIS25k .
We construct the first asymptotically good relaxed locally correctable codes with polylogarithmic query complexity, bringing the upper bound polynomially close to the lower bound of Gur and Lachish (SICOMP 2021). Our result follows from showing that a high-rate locally testable code can boost the block length of a smaller relaxed locally correctable code, while preserving the correcting radius and incurring only a modest additive cost in rate and query complexity. We use the locally testable code's tester to check if the amount of corruption in the input is low; if so, we can "zoom-in" to a suitable substring of the input and recurse on the smaller code's local corrector. Hence, iterating this operation with a suitable family of locally testable codes due to Dinur, Evra, Livne, Lubotzky, and Mozes (STOC 2022) yields asymptotically good codes with relaxed local correctability, arbitrarily large block length, and polylogarithmic query complexity. Our codes asymptotically inherit the rate and distance of any locally testable code used in the final invocation of the operation. Therefore, our framework also yields nonexplicit relaxed locally correctable codes with polylogarithmic query complexity that have rate and distance approaching the Gilbert-Varshamov bound.
High spectral dimensionality and the shortage of annotations make hyperspectral image (HSI) classification a challenging problem. Recent studies suggest that convolutional neural networks can learn discriminative spatial features, which play a paramount role in HSI interpretation. However, most of these methods ignore the distinctive spectral-spatial characteristic of hyperspectral data. In addition, a large amount of unlabeled data remains an unexploited gold mine for efficient data use. Therefore, we proposed an integration of generative adversarial networks (GANs) and probabilistic graphical models for HSI classification. Specifically, we used a spectral-spatial generator and a discriminator to identify land cover categories of hyperspectral cubes. Moreover, to take advantage of a large amount of unlabeled data, we adopted a conditional random field to refine the preliminary classification results generated by GANs. Experimental results obtained using two commonly studied datasets demonstrate that the proposed framework achieved encouraging classification accuracy using a small number of data for training.
注冊地址： 北京市海淀區羊坊店路18號2幢3層301-191