Node classification on graph data is a major problem, and various graph neural networks (GNNs) have been proposed. Variants of GNNs such as H2GCN and CPF outperform graph convolutional networks (GCNs) by improving on the weaknesses of the traditional GNN. However, there are some graph data which these GNN variants fail to perform well than other GNNs in the node classification task. This is because H2GCN has a feature thinning on graph data with high average degree, and CPF gives rise to a problem about label-propagation suitability. Accordingly, we propose a hierarchical model selection framework (HMSF) that selects an appropriate GNN model by analyzing the indicators of each graph data. In the experiment, we show that the model selected by our HMSF achieves high performance on node classification for various types of graph data.
This white paper lays out a vision of research and development in the field of artificial intelligence for the next decade (and beyond). Its denouement is a cyber-physical ecosystem of natural and synthetic sense-making, in which humans are integral participants -- what we call ''shared intelligence''. This vision is premised on active inference, a formulation of adaptive behavior that can be read as a physics of intelligence, and which inherits from the physics of self-organization. In this context, we understand intelligence as the capacity to accumulate evidence for a generative model of one's sensed world -- also known as self-evidencing. Formally, this corresponds to maximizing (Bayesian) model evidence, via belief updating over several scales: i.e., inference, learning, and model selection. Operationally, this self-evidencing can be realized via (variational) message passing or belief propagation on a factor graph. Crucially, active inference foregrounds an existential imperative of intelligent systems; namely, curiosity or the resolution of uncertainty. This same imperative underwrites belief sharing in ensembles of agents, in which certain aspects (i.e., factors) of each agent's generative world model provide a common ground or frame of reference. Active inference plays a foundational role in this ecology of belief sharing -- leading to a formal account of collective intelligence that rests on shared narratives and goals. We also consider the kinds of communication protocols that must be developed to enable such an ecosystem of intelligences and motivate the development of a shared hyper-spatial modeling language and transaction protocol, as a first -- and key -- step towards such an ecology.
Out-Of-Distribution (OOD) detection has received broad attention over the years, aiming to ensure the reliability and safety of deep neural networks (DNNs) in real-world scenarios by rejecting incorrect predictions. However, we notice a discrepancy between the conventional evaluation vs. the essential purpose of OOD detection. On the one hand, the conventional evaluation exclusively considers risks caused by label-space distribution shifts while ignoring the risks from input-space distribution shifts. On the other hand, the conventional evaluation reward detection methods for not rejecting the misclassified image in the validation dataset. However, the misclassified image can also cause risks and should be rejected. We appeal to rethink OOD detection from a human-centric perspective, that a proper detection method should reject the case that the deep model's prediction mismatches the human expectations and adopt the case that the deep model's prediction meets the human expectations. We propose a human-centric evaluation and conduct extensive experiments on 45 classifiers and 8 test datasets. We find that the simple baseline OOD detection method can achieve comparable and even better performance than the recently proposed methods, which means that the development in OOD detection in the past years may be overestimated. Additionally, our experiments demonstrate that model selection is non-trivial for OOD detection and should be considered as an integral of the proposed method, which differs from the claim in existing works that proposed methods are universal across different models.
When several models have similar training scores, classical model selection heuristics follow Occam's razor and advise choosing the ones with least capacity. Yet, modern practice with large neural networks has often led to situations where two networks with exactly the same number of parameters score similar on the training set, but the deeper one generalizes better to unseen examples. With this in mind, it is well accepted that deep networks are superior to shallow wide ones. However, theoretically there is no difference between the two. In fact, they are both universal approximators. In this work we propose a new unsupervised measure that predicts how well a model will generalize. We call it the output sharpness, and it is based on the fact that, in reality, boundaries between concepts are generally unsharp. We test this new measure on several neural network settings, and architectures, and show how generally strong the correlation is between our metric, and test set performance. Having established this measure, we give a mathematical probabilistic argument that predicts network depth to be correlated with our proposed measure. After verifying this in real data, we are able to formulate the key argument of the work: output sharpness hampers generalization; deep networks have an in built bias against it; therefore, deep networks beat wide ones. All in all the work not only provides a helpful predictor of overfitting that can be used in practice for model selection (or even regularization), but also provides a much needed theoretical grounding for the success of modern deep neural networks.
We study the problem of out-of-distribution (o.o.d.) generalization where spurious correlations of attributes vary across training and test domains. This is known as the problem of correlation shift and has posed concerns on the reliability of machine learning. In this work, we introduce the concepts of direct and indirect effects from causal inference to the domain generalization problem. We argue that models that learn direct effects minimize the worst-case risk across correlation-shifted domains. To eliminate the indirect effects, our algorithm consists of two stages: in the first stage, we learn an indirect-effect representation by minimizing the prediction error of domain labels using the representation and the class labels; in the second stage, we remove the indirect effects learned in the first stage by matching each data with another data of similar indirect-effect representation but of different class labels in the training and validation phase. Our approach is shown to be compatible with existing methods and improve the generalization performance of them on correlation-shifted datasets. Experiments on 5 correlation-shifted datasets and the DomainBed benchmark verify the effectiveness of our approach.
Selecting the right set of hyperparameters is crucial in time series forecasting. The classical temporal cross-validation framework for hyperparameter optimization (HPO) often leads to poor test performance because of a possible mismatch between validation and test periods. To address this test-validation mismatch, we propose a novel technique, H-Pro to drive HPO via test proxies by exploiting data hierarchies often associated with time series datasets. Since higher-level aggregated time series often show less irregularity and better predictability as compared to the lowest-level time series which can be sparse and intermittent, we optimize the hyperparameters of the lowest-level base-forecaster by leveraging the proxy forecasts for the test period generated from the forecasters at higher levels. H-Pro can be applied on any off-the-shelf machine learning model to perform HPO. We validate the efficacy of our technique with extensive empirical evaluation on five publicly available hierarchical forecasting datasets. Our approach outperforms existing state-of-the-art methods in Tourism, Wiki, and Traffic datasets, and achieves competitive result in Tourism-L dataset, without any model-specific enhancements. Moreover, our method outperforms the winning method of the M5 forecast accuracy competition.
A biological system is a complex network of heterogeneous molecular entities and their interactions contributing to various biological characteristics of the system. However, current biological networks are noisy, sparse, and incomplete, limiting our ability to create a holistic view of the biological system and understand the biological phenomena. Experimental identification of such interactions is both time-consuming and expensive. With the recent advancements in high-throughput data generation and significant improvement in computational power, various computational methods have been developed to predict novel interactions in the noisy network. Recently, deep learning methods such as graph neural networks have shown their effectiveness in modeling graph-structured data and achieved good performance in biomedical interaction prediction. However, graph neural networks-based methods require human expertise and experimentation to design the appropriate complexity of the model and significantly impact the performance of the model. Furthermore, deep graph neural networks face overfitting problems and tend to be poorly calibrated with high confidence on incorrect predictions. To address these challenges, we propose Bayesian model selection for graph convolutional networks to jointly infer the most plausible number of graph convolution layers (depth) warranted by data and perform dropout regularization simultaneously. Experiments on four interaction datasets show that our proposed method achieves accurate and calibrated predictions. Our proposed method enables the graph convolutional networks to dynamically adapt their depths to accommodate an increasing number of interactions.
Plotting a learner's generalization performance against the training set size results in a so-called learning curve. This tool, providing insight in the behavior of the learner, is also practically valuable for model selection, predicting the effect of more training data, and reducing the computational complexity of training. We set out to make the (ideal) learning curve concept precise and briefly discuss the aforementioned usages of such curves. The larger part of this survey's focus, however, is on learning curves that show that more data does not necessarily leads to better generalization performance. A result that seems surprising to many researchers in the field of artificial intelligence. We point out the significance of these findings and conclude our survey with an overview and discussion of open problems in this area that warrant further theoretical and empirical investigation.
For time-critical IoT applications using deep learning, inference acceleration through distributed computing is a promising approach to meet a stringent deadline. In this paper, we implement a working prototype of a new distributed inference acceleration method HALP using three raspberry Pi 4. HALP accelerates inference by designing a seamless collaboration among edge devices (EDs) in Edge Computing. We maximize the parallelization between communication and computation among the collaborative EDs by optimizing the task partitioning ratio based on the segment-based partitioning. Experimental results show that the distributed inference HALP achieves 1.7x inference acceleration for VGG-16. Then, we combine distributed inference with conventional neural network model compression by setting up different shrinking hyperparameters for MobileNet-V1. In this way, we can further accelerate inference but at the cost of inference accuracy loss. To strike a balance between latency and accuracy, we propose dynamic model selection to select a model which provides the highest accuracy within the latency constraint. It is shown that the model selection with distributed inference HALP can significantly improve service reliability compared to the conventional stand-alone computation.
The well-documented presence of texture bias in modern convolutional neural networks has led to a plethora of algorithms that promote an emphasis on shape cues, often to support generalization to new domains. Yet, common datasets, benchmarks and general model selection strategies are missing, and there is no agreed, rigorous evaluation protocol. In this paper, we investigate difficulties and limitations when training networks with reduced texture bias. In particular, we also show that proper evaluation and meaningful comparisons between methods are not trivial. We introduce BiasBed, a testbed for texture- and style-biased training, including multiple datasets and a range of existing algorithms. It comes with an extensive evaluation protocol that includes rigorous hypothesis testing to gauge the significance of the results, despite the considerable training instability of some style bias methods. Our extensive experiments, shed new light on the need for careful, statistically founded evaluation protocols for style bias (and beyond). E.g., we find that some algorithms proposed in the literature do not significantly mitigate the impact of style bias at all. With the release of BiasBed, we hope to foster a common understanding of consistent and meaningful comparisons, and consequently faster progress towards learning methods free of texture bias. Code is available at https://github.com/D1noFuzi/BiasBed
Deep learning has been widely used in various applications from different fields such as computer vision, natural language processing, etc. However, the training models are often manually developed via many costly experiments. This manual work usually requires substantial computing resources, time, and experience. To simplify the use of deep learning and alleviate human effort, automated deep learning has emerged as a potential tool that releases the burden for both users and researchers. Generally, an automatic approach should support the diversity of model selection and the evaluation should allow users to decide upon their demands. To that end, we propose a multi-objective end-to-end self-optimized approach for constructing deep learning models automatically. Experimental results on well-known datasets such as MNIST, Fashion, and Cifar10 show that our algorithm can discover various competitive models compared with the state-of-the-art approach. In addition, our approach also introduces multi-objective trade-off solutions for both accuracy and uncertainty metrics for users to make better decisions.
In model-based optimisation (MBO) we are interested in using machine learning to design candidates that maximise some measure of reward with respect to a black box function called the (ground truth) oracle, which is expensive to compute since it involves executing a real world process. In offline MBO we wish to do so without assuming access to such an oracle during training or validation, with makes evaluation non-straightforward. While an approximation to the ground oracle can be trained and used in place of it during model validation to measure the mean reward over generated candidates, the evaluation is approximate and vulnerable to adversarial examples. Measuring the mean reward of generated candidates over this approximation is one such `validation metric', whereas we are interested in a more fundamental question which is finding which validation metrics correlate the most with the ground truth. This involves proposing validation metrics and quantifying them over many datasets for which the ground truth is known, for instance simulated environments. This is encapsulated under our proposed evaluation framework which is also designed to measure extrapolation, which is the ultimate goal behind leveraging generative models for MBO. While our evaluation framework is model agnostic we specifically evaluate denoising diffusion models due to their state-of-the-art performance, as well as derive interesting insights such as ranking the most effective validation metrics as well as discussing important hyperparameters.
Chain event graphs (CEGs) are a recent family of probabilistic graphical models that generalise the popular Bayesian networks (BNs) family. Crucially, unlike BNs, a CEG is able to embed, within its graph and its statistical model, asymmetries exhibited by a process. These asymmetries might be in the conditional independence relationships or in the structure of the graph and its underlying event space. Structural asymmetries are common in many domains, and can occur naturally (e.g. a defendant vs prosecutor's version of events) or by design (e.g. a public health intervention). However, there currently exists no software that allows a user to leverage the theoretical developments of the CEG model family in modelling processes with structural asymmetries. This paper introduces cegpy, the first Python package for learning and analysing complex processes using CEGs. The key feature of cegpy is that it is the first CEG package in any programming language that can model processes with symmetric as well as asymmetric structures. cegpy contains an implementation of Bayesian model selection and probability propagation algorithms for CEGs. We illustrate the functionality of cegpy using a structurally asymmetric dataset.
The theory of bias-variance used to serve as a guide for model selection when applying Machine Learning algorithms. However, modern practice has shown success with over-parameterized models that were expected to overfit but did not. This led to the proposal of the double descent curve of performance by Belkin et al. Although it seems to describe a real, representative phenomenon, the field is lacking a fundamental theoretical understanding of what is happening, what are the consequences for model selection and when is double descent expected to occur. In this paper we develop a principled understanding of the phenomenon, and sketch answers to these important questions. Furthermore, we report real experimental results that are correctly predicted by our proposed hypothesis.
Signed networks, i.e., networks with positive and negative edges, commonly arise in various domains from social media to epidemiology. Modeling signed networks has many practical applications, including the creation of synthetic data sets for experiments where obtaining real data is difficult. Influential prior works proposed and studied various graph topology models, as well as the problem of selecting the most fitting model for different application domains. However, these topology models are typically unsigned. In this work, we pose a novel Maximum-Likelihood-based optimization problem for modeling signed networks given their topology and showcase it in the context of gene regulation. Regulatory interactions of genes play a key role in organism development, and when broken can lead to serious organism abnormalities and diseases. Our contributions are threefold: First, we design a new class of signage models for a given topology. Based on the parameter setting, we discuss its biological interpretations for gene regulatory networks (GRNs). Second, we design algorithms computing the Maximum Likelihood -- depending on the parameter setting, our algorithms range from closed-form expressions to MCMC sampling. Third, we evaluated the results of our algorithms on synthetic datasets and real-world large GRNs. Our work can lead to the prediction of unknown gene regulations, the generation of biological hypotheses, and realistic GRN benchmark datasets.
Code generation models can benefit data scientists' productivity by automatically generating code from context and text descriptions. An important measure of the modeling progress is whether a model can generate code that can correctly execute to solve the task. However, due to the lack of an evaluation dataset that directly supports execution-based model evaluation, existing work relies on code surface form similarity metrics (e.g., BLEU, CodeBLEU) for model selection, which can be inaccurate. To remedy this, we introduce ExeDS, an evaluation dataset for execution evaluation for data science code generation tasks. ExeDS contains a set of 534 problems from Jupyter Notebooks, each consisting of code context, task description, reference program, and the desired execution output. With ExeDS, we evaluate the execution performance of five state-of-the-art code generation models that have achieved high surface-form evaluation scores. Our experiments show that models with high surface-form scores do not necessarily perform well on execution metrics, and execution-based metrics can better capture model code generation errors. Source code and data can be found at https://github.com/Jun-jie-Huang/ExeDS
Similar activity patterns may arise from model neural networks with distinct coupling properties and individual unit dynamics. These similar patterns may, however, respond differently to parameter variations and, specifically, to tuning of inputs that represent control signals. In this work, we analyze the responses resulting from modulation of a localized input in each of three classes of model neural networks that have been recognized in the literature for their capacity to produce robust three-phase rhythms: coupled fast-slow oscillators, near-heteroclinic oscillators, and threshold-linear networks. Triphasic rhythms, in which each phase consists of a prolonged activation of a corresponding subgroup of neurons followed by a fast transition to another phase, represent a fundamental activity pattern observed across a range of central pattern generators underlying behaviors critical to survival, including respiration, locomotion, and feeding. To perform our analysis, we extend the recently developed local timing response curve (lTRC), which allows us to characterize the timing effects due to perturbations, and we complement our lTRC approach with model-specific dynamical systems analysis. Interestingly, we observe disparate effects of similar perturbations across distinct model classes. Thus, this work provides an analytical framework for studying control of oscillations in nonlinear dynamical systems, and may help guide model selection in future efforts to study systems exhibiting triphasic rhythmic activity.
Forecasts of regional electricity net-demand, consumption minus embedded generation, are an essential input for reliable and economic power system operation, and energy trading. While such forecasts are typically performed region by region, operations such as managing power flows require spatially coherent joint forecasts, which account for cross-regional dependencies. Here, we forecast the joint distribution of net-demand across the 14 regions constituting Great Britain's electricity network. Joint modelling is complicated by the fact that the net-demand variability within each region, and the dependencies between regions, vary with temporal, socio-economical and weather-related factors. We accommodate for these characteristics by proposing a multivariate Gaussian model based on a modified Cholesky parametrisation, which allows us to model each unconstrained parameter via an additive model. Given that the number of model parameters and covariates is large, we adopt a semi-automated approach to model selection, based on gradient boosting. In addition to comparing the forecasting performance of several versions of the proposed model with that of two non-Gaussian copula-based models, we visually explore the model output to interpret how the covariates affect net-demand variability and dependencies. The code for reproducing the results in this paper is available at https://doi.org/10.5281/zenodo.7315105, while methods for building and fitting multivariate Gaussian additive models are provided by the SCM R package, available at https://github.com/VinGioia90/SCM.
Community detection is the task of discovering groups of nodes sharing similar patterns within a network. With recent advancements in deep learning, methods utilizing graph representation learning and deep clustering have shown great results in community detection. However, these methods often rely on the topology of networks (i) ignoring important features such as network heterogeneity, temporality, multimodality, and other possibly relevant features. Besides, (ii) the number of communities is not known a priori and is often left to model selection. In addition, (iii) in multimodal networks all nodes are assumed to be symmetrical in their features; while true for homogeneous networks, most of the real-world networks are heterogeneous where feature availability often varies. In this paper, we propose a novel framework (named MGTCOM) that overcomes the above challenges (i)--(iii). MGTCOM identifies communities through multimodal feature learning by leveraging a new sampling technique for unsupervised learning of temporal embeddings. Importantly, MGTCOM is an end-to-end framework optimizing network embeddings, communities, and the number of communities in tandem. In order to assess its performance, we carried out an extensive evaluation on a number of multimodal networks. We found out that our method is competitive against state-of-the-art and performs well in inductive inference.
In Gaussian graphical model selection, noise-corrupted samples present significant challenges. It is known that even minimal amounts of noise can obscure the underlying structure, leading to fundamental identifiability issues. A recent line of work addressing this "robust model selection" problem narrows its focus to tree-structured graphical models. Even within this specific class of models, exact structure recovery is shown to be impossible. However, several algorithms have been developed that are known to provably recover the underlying tree-structure up to an (unavoidable) equivalence class. In this paper, we extend these results beyond tree-structured graphs. We first characterize the equivalence class up to which general graphs can be recovered in the presence of noise. Despite the inherent ambiguity (which we prove is unavoidable), the structure that can be recovered reveals local clustering information and global connectivity patterns in the underlying model. Such information is useful in a range of real-world problems, including power grids, social networks, protein-protein interactions, and neural structures. We then propose an algorithm which provably recovers the underlying graph up to the identified ambiguity. We further provide finite sample guarantees in the high-dimensional regime for our algorithm and validate our results through numerical simulations.