Crowd sequential annotations can be an efficient and cost-effective way to build large datasets for sequence labeling. Different from tagging independent instances, for crowd sequential annotations the quality of label sequence relies on the expertise level of annotators in capturing internal dependencies for each token in the sequence. In this paper, we propose Modeling sequential annotation for sequence labeling with crowds (SA-SLC). First, a conditional probabilistic model is developed to jointly model sequential data and annotators' expertise, in which categorical distribution is introduced to estimate the reliability of each annotator in capturing local and non-local label dependency for sequential annotation. To accelerate the marginalization of the proposed model, a valid label sequence inference (VLSE) method is proposed to derive the valid ground-truth label sequences from crowd sequential annotations. VLSE derives possible ground-truth labels from the token-wise level and further prunes sub-paths in the forward inference for label sequence decoding. VLSE reduces the number of candidate label sequences and improves the quality of possible ground-truth label sequences. The experimental results on several sequence labeling tasks of Natural Language Processing show the effectiveness of the proposed model.

There has been a growing interest in Machine Unlearning recently, primarily due to legal requirements such as the General Data Protection Regulation (GDPR) and the California Consumer Privacy Act. Thus, multiple approaches were presented to remove the influence of specific target data points from a trained model. However, when evaluating the success of unlearning, current approaches either use adversarial attacks or compare their results to the optimal solution, which usually incorporates retraining from scratch. We argue that both ways are insufficient in practice. In this work, we present an evaluation metric for Machine Unlearning algorithms based on epistemic uncertainty. This is the first definition of a general evaluation metric for Machine Unlearning to our best knowledge.

In this paper, we propose a new covering technique localized for the trajectories of SGD. This localization provides an algorithm-specific complexity measured by the covering number, which can have dimension-independent cardinality in contrast to standard uniform covering arguments that result in exponential dimension dependency. Based on this localized construction, we show that if the objective function is a finite perturbation of a piecewise strongly convex and smooth function with $P$ pieces, i.e. non-convex and non-smooth in general, the generalization error can be upper bounded by $O(\sqrt{(\log n\log(nP))/n})$, where $n$ is the number of data samples. In particular, this rate is independent of dimension and does not require early stopping and decaying step size. Finally, we employ these results in various contexts and derive generalization bounds for multi-index linear models, multi-class support vector machines, and $K$-means clustering for both hard and soft label setups, improving the known state-of-the-art rates.

Neural networks have achieved remarkable success in solving large-scale machine learning problems in various application domains such as computer vision and natural language processing [33]. Firstorder methods such as gradient descent (GD) and stochastic gradient descent (SGD) are mainstream optimization algorithms for training neural networks due to their simplicity and efficiency [11, 33, 50]. Although the associated optimization problems are nonconvex and nonsmooth, GD/SGD can still find a model with a very small or even zero training error [16, 20, 34, 39, 64, 69]. At the same time, the models found by such first-order methods has demonstrated good generalization performance on test data despite neural networks are often highly overparameterized in the sense that the number of parameters is much larger than the size of training examples [1, 2, 5]. These surprising phenomena have triggered a surge of research activities in understanding the generalization ability of neural networks. Generalization analysis typically uses complexity measures such as VC dimension, covering numbers or Rademacher complexities to develop capacity-dependent bounds [8, 9, 25, 42, 48], which, however, may not explain well the generalization of overparameterized neural networks. Impressive alternatives have been proposed which include the compression approach [4], the norm-based analysis [8, 25], the PAC-Bayes analysis [21] and the neural tangent kernel (NTK) approach [5, 28]. In particular, the NTK approach shows that the overparameterization pulls the dynamic of GD on neural networks close to its counterpart on a kernelized machine with the least-square loss [5, 20], which shows how overparameterization can help both optimization and generalization. However, this approach often requires a very high overparameterization to gain useful results [6, 55, 60].

We investigate a novel non-parametric regression-based clustering algorithm for longitudinal data analysis. Combining natural cubic splines with Gaussian mixture models (GMM), the algorithm can produce smooth cluster means that describe the underlying data well. However, there are some shortcomings in the algorithm: high computational complexity in the parameter estimation procedure and a numerically unstable variance estimator. Therefore, to further increase the usability of the method, we incorporated approaches to reduce its computational complexity, we developed a new, more stable variance estimator, and we developed a new smoothing parameter estimation procedure. We show that the developed algorithm, SMIXS, performs better than GMM on a synthetic dataset in terms of clustering and regression performance. We demonstrate the impact of the computational speed-ups, which we formally prove in the new framework. Finally, we perform a case study by using SMIXS to cluster vertical atmospheric measurements to determine different weather regimes.

If the trend of learned components eventually outperforming their hand-crafted version continues, learned optimizers will eventually outperform hand-crafted optimizers like SGD or Adam. Even if learned optimizers (L2Os) eventually outpace hand-crafted ones in practice however, they are still not provably convergent and might fail out of distribution. These are the questions addressed here. Currently, learned optimizers frequently outperform generic hand-crafted optimizers (such as gradient descent) at the beginning of learning but they generally plateau after some time while the generic algorithms continue to make progress and often overtake the learned algorithm as Aesop's tortoise which overtakes the hare. L2Os also still have a difficult time generalizing out of distribution. Heaton et al. proposed Safeguarded L2O (GL2O) which can take a learned optimizer and safeguard it with a generic learning algorithm so that by conditionally switching between the two, the resulting algorithm is provably convergent. We propose a new class of Safeguarded L2O, called Loss-Guarded L2O (LGL2O), which is both conceptually simpler and computationally less expensive. The guarding mechanism decides solely based on the expected future loss value of both optimizers. Furthermore, we show theoretical proof of LGL2O's convergence guarantee and empirical results comparing to GL2O and other baselines showing that it combines the best of both L2O and SGD and that in practice converges much better than GL2O.

In the last few years, several initiatives have been starting to offer access to research outputs data and metadata in an open fashion. The platforms developed by those initiatives are opening up scientific production to the wider public and they can be an invaluable asset for evidence-based policy-making in Science, Technology and Innovation (STI). These resources can indeed facilitate knowledge discovery and help identify available R&D assets and relevant actors within specific research niches of interest. Ideally, to gain a comprehensive view of entire STI ecosystems, the information provided by each of these resources should be combined and analysed accordingly. To ensure so, at least a certain degree of interoperability should be guaranteed across data sources, so that data could be better aggregated and complemented and that evidence provided towards policy-making is more complete and reliable. Here, we study whether this is the case for the case of mapping Climate Action research in the whole Denmark STI ecosystem, by using 4 popular open access STI data sources, namely OpenAire, Open Alex, CORDIS and Kohesio.

This paper examines two different yet related questions related to explainable AI (XAI) practices. Machine learning (ML) is increasingly important in financial services, such as pre-approval, credit underwriting, investments, and various front-end and back-end activities. Machine Learning can automatically detect non-linearities and interactions in training data, facilitating faster and more accurate credit decisions. However, machine learning models are opaque and hard to explain, which are critical elements needed for establishing a reliable technology. The study compares various machine learning models, including single classifiers (logistic regression, decision trees, LDA, QDA), heterogeneous ensembles (AdaBoost, Random Forest), and sequential neural networks. The results indicate that ensemble classifiers and neural networks outperform. In addition, two advanced post-hoc model agnostic explainability techniques - LIME and SHAP are utilized to assess ML-based credit scoring models using the open-access datasets offered by US-based P2P Lending Platform, Lending Club. For this study, we are also using machine learning algorithms to develop new investment models and explore portfolio strategies that can maximize profitability while minimizing risk.

In semiconductor manufacturing, wafer map defect pattern provides critical information for facility maintenance and yield management, so the classification of defect patterns is one of the most important tasks in the manufacturing process. In this paper, we propose a novel way to represent the shape of the defect pattern as a finite-dimensional vector, which will be used as an input for a neural network algorithm for classification. The main idea is to extract the topological features of each pattern by using the theory of persistent homology from topological data analysis (TDA). Through some experiments with a simulated dataset, we show that the proposed method is faster and much more efficient in training with higher accuracy, compared with the method using convolutional neural networks (CNN) which is the most common approach for wafer map defect pattern classification. Moreover, our method outperforms the CNN-based method when the number of training data is not enough and is imbalanced.

Function plays an important role in mathematics and many science branches. As the fast development of computer technology, more and more study on computational function analysis, e.g., Fast Fourier Transform, Wavelet Transform, Curve Function, are presented in these years. However, there are two main problems in these approaches: 1) hard to handle the complex functions of stationary and non-stationary, periodic and non-periodic, high order and low order; 2) hard to generalize the fitting functions from training data to test data. In this paper, a multiple regression based function fitting network that solves the two main problems is introduced as a predicable function fitting technique. This technique constructs the network includes three main parts: 1) the stationary transform layer, 2) the feature encoding layers, and 3) the fine tuning regression layer. The stationary transform layer recognizes the order of input function data, and transforms non-stationary function to stationary function. The feature encoding layers encode the raw input sequential data to a novel linear regression feature that can capture both the structural and the temporal characters of the sequential data. The fine tuning regression layer then fits the features to the target ahead values. The fitting network with the linear regression feature layers and a non-linear regression layer come up with high quality fitting results and generalizable predictions. The experiments of both mathematic function examples and the real word function examples verifies the efficiency of the proposed technique.