Most current methods for learning from demonstrations assume that those demonstrations alone are sufficient to learn the underlying task. This is often untrue, especially if extra safety specifications exist which were not present in the original demonstrations. In this paper, we allow an expert to elaborate on their original demonstration with additional specification information using linear temporal logic (LTL). Our system converts LTL specifications into a differentiable loss. This loss is then used to learn a dynamic movement primitive that satisfies the underlying specification, while remaining close to the original demonstration. Further, by leveraging adversarial training, our system learns to robustly satisfy the given LTL specification on unseen inputs, not just those seen in training. We show that our method is expressive enough to work across a variety of common movement specification patterns such as obstacle avoidance, patrolling, keeping steady, and speed limitation. In addition, we show that our system can modify a base demonstration with complex specifications by incrementally composing multiple simpler specifications. We also implement our system on a PR-2 robot to show how a demonstrator can start with an initial (sub-optimal) demonstration, then interactively improve task success by including additional specifications enforced with our differentiable LTL loss.

Chest radiographs are commonly performed low-cost exams for screening and diagnosis. However, radiographs are 2D representations of 3D structures causing considerable clutter impeding visual inspection and automated image analysis. Here, we propose a Fully Convolutional Network to suppress, for a specific task, undesired visual structure from radiographs while retaining the relevant image information such as lung-parenchyma. The proposed algorithm creates reconstructed radiographs and ground-truth data from high resolution CT-scans. Results show that removing visual variation that is irrelevant for a classification task improves the performance of a classifier when only limited training data are available. This is particularly relevant because a low number of ground-truth cases is common in medical imaging.

Graphs are versatile tools for representing structured data. Therefore, a variety of machine learning methods have been studied for graph data analysis. Although many of those learning methods depend on the measurement of differences between input graphs, defining an appropriate distance metric for a graph remains a controversial issue. Hence, we propose a supervised distance metric learning method for the graph classification problem. Our method, named interpretable graph metric learning (IGML), learns discriminative metrics in a subgraph-based feature space, which has a strong graph representation capability. By introducing a sparsity-inducing penalty on a weight of each subgraph, IGML can identify a small number of important subgraphs that can provide insight about the given classification task. Since our formulation has a large number of optimization variables, an efficient algorithm is also proposed by using pruning techniques based on safe screening and working set selection methods. An important property of IGML is that the optimality of the solution is guaranteed because the problem is formulated as a convex problem and our pruning strategies only discard unnecessary subgraphs. Further, we show that IGML is also applicable to other structured data such as item-set and sequence data, and that it can incorporate vertex-label similarity by using a transportation-based subgraph feature. We empirically evaluate the computational efficiency and classification performance on several benchmark datasets and show some illustrative examples demonstrating that IGML identifies important subgraphs from a given graph dataset.

Most decision tree induction algorithms are based on a greedy top-down recursive partitioning strategy for tree growth. In this paper, we propose several methods for induction of decision trees and their ensembles based on evolutionary algorithms. The main difference of our approach is using real-valued vector representation of decision tree that allows to use a large number of different optimization algorithms, as well as optimize the whole tree or ensemble for avoiding local optima. Differential evolution and evolution strategies were chosen as optimization algorithms, as they have good results in reinforcement learning problems. We test the predictive performance of this methods using several public UCI data sets, and the proposed methods show better quality than classical methods.

We consider building predictors when the data have missing values. We study the seemingly-simple case where the target to predict is a linear function of the fully-observed data and we show that, in the presence of missing values, the optimal predictor may not be linear. In the particular Gaussian case, it can be written as a linear function of multiway interactions between the observed data and the various missing-value indicators. Due to its intrinsic complexity, we study a simple approximation and prove generalization bounds with finite samples, highlighting regimes for which each method performs best. We then show that multilayer perceptrons with ReLU activation functions can be consistent, and can explore good trade-offs between the true model and approximations. Our study highlights the interesting family of models that are beneficial to fit with missing values depending on the amount of data available.

Maintaining a population of solutions has been shown to increase exploration in reinforcement learning, typically attributed to the greater diversity of behaviors considered. One such class of methods, novelty search, considers further boosting diversity across agents via a multi-objective optimization formulation. Despite the intuitive appeal, these mechanisms have several shortcomings. First, they make use of mean field updates, which induce cycling behaviors. Second, they often rely on handcrafted behavior characterizations, which require domain knowledge. Furthermore, boosting diversity often has a detrimental impact on optimizing already fruitful behaviors for rewards. Setting the relative importance of novelty- versus reward-factor is usually hardcoded or requires tedious tuning/annealing. In this paper, we introduce a novel measure of population-wide diversity, leveraging ideas from Determinantal Point Processes. We combine this in a principled fashion with the reward function to adapt to the degree of diversity during training, borrowing ideas from online learning. Combined with task-agnostic behavioral embeddings, we show this approach outperforms previous methods for multi-objective optimization, as well as vanilla algorithms solely optimizing for rewards.

Neural network pruning is a popular method to reduce the size of a trained model, allowing efficient computation during inference time, with minimal loss in accura cy. However, such a method still requires the process of training an over-parameterized network, as trai ning a pruned network from scratch seems to fail (see [ 10 ]). Recently, a work by Frankle and Carbin [ 10 ] has presented a surprising phenomenon: pruned neural networks can be trained to achieve good performance, when resetting their weights to their initial values. Hence, the authors state the lottery ticket hypothesis: a randomly-initialized neural network contains a subnetwork such that, when trained in isolation, can match the performance of the original network. This observation has attracted great interest, with variou s followup works trying to understand this intriguing phenomenon. Specifically, very recent works by Z hou et al. [ 37 ], Ramanujan et al. [ 27 ] presented algorithms to find subnetworks that already achieve good per formance, without any training.

--Federated Learning (FL) is viewed as a promising technique for future distributed machine learning. It permits a large number of mobile devices participating in the training of a global model collaboratively without having to expose their local private data. Although the challenge of the network connection will be much relieved in 5G/B5G era, the training latency is still an obstacle preventing FL from being largely adopted. One of the most fundamental problems that leads to large training latency is the bad candidate-selection of FL participants. T o the best of our knowledge, the existing candidate-selection algorithms belong to the reactive manner . Under such reactive selection, the FL parameter server only knows the currently-observed resources of all candidates. In the dynamic FL environment, the mobile devices selected by the reactive candidate-selection algorithms very possibly fail to complete the training and reporting phases of FL. T o this end, we study the proactive candidate-selection for FL in this paper . We first let each candidate device locally predict the qualities of both its training and reporting phases using the LSTM network. Then, the proposed candidate-selection algorithm is implemented by the Deep Reinforcement Learning (DRL) framework, which can adapt to the dynamically varying factors in the metropolitan edge computing environment. Finally, the real-world trace-driven experiments prove that the proposed proactive approach outperforms the existing reactive algorithms with respect to the ratio of valid participants and the test accuracy of the aggregated global FL model. Federated Learning (FL) [1], [2] is a branch of distributed machine learning that enables a group of distributed devices to train their individual local models using the local dataset. Thus, FL is a promising computing paradigm in our future intelligent life, especially under the fifth generation (5G) and the beyond (B5G) communications networks. For example, the FederatedAveraging (FedAvg) algorithm [1] can help mobile users predict the next-words when users are using the Google's GBoard [3] in their smartphones. To realize a large-scale federated learning framework, a number of challenges must be addressed.

Recent years have witnessed an abundance of new publications and approaches on meta-learning. This community-wide enthusiasm has sparked great insights but has also created a plethora of seemingly different frameworks, which can be hard to compare and evaluate. In this paper, we aim to provide a principled, unifying framework by revisiting and strengthening the connection between meta-learning and traditional supervised learning. By treating pairs of task-specific data sets and target models as (feature, label) samples, we can reduce many meta-learning algorithms to instances of supervised learning. This view not only unifies meta-learning into an intuitive and practical framework but also allows us to transfer insights from supervised learning directly to improve meta-learning. For example, we obtain a better understanding of generalization properties, and we can readily transfer well-understood techniques, such as model ensemble, pre-training, joint training, data augmentation, and even nearest neighbor based methods. We provide an intuitive analogy of these methods in the context of meta-learning and show that they give rise to significant improvements in model performance on few-shot learning.

We present the implementation of a new stochastic Kuramoto-Vicsek-type model for global optimization of nonconvex functions on the sphere. This model belongs to the class of Consensus-Based Optimization. In fact, particles move on the sphere driven by a drift towards an instantaneous consensus point, which is computed as a convex combination of particle locations, weighted by the cost function according to Laplace's principle, and it represents an approximation to a global minimizer. The dynamics is further perturbed by a random vector field to favor exploration, whose variance is a function of the distance of the particles to the consensus point. In particular, as soon as the consensus is reached the stochastic component vanishes. The main results of this paper are about the proof of convergence of the numerical scheme to global minimizers provided conditions of well-preparation of the initial datum. The proof combines previous results of mean-field limit with a novel asymptotic analysis, and classical convergence results of numerical methods for SDE. We present several numerical experiments, which show that the algorithm proposed in the present paper scales well with the dimension and is extremely versatile. To quantify the performances of the new approach, we show that the algorithm is able to perform essentially as good as ad hoc state of the art methods in challenging problems in signal processing and machine learning, namely the phase retrieval problem and the robust subspace detection.