arXiv subjects

Tobias Glasmachers

Publications and source records attributed to Tobias Glasmachers.

At least 37 records · Page 2Linked to original sources

arXiv2023

Understanding Activation Patterns in Artificial Neural Networks by Exploring Stochastic Processes

To gain a deeper understanding of the behavior and learning dynamics of (deep) artificial neural networks, it is valuable to employ mathematical abstractions and models. These tools provide a simplified perspective on network performance and facilitate systematic investigations through simulations. In this paper, we propose utilizing the framework of stochastic processes, which has been underutilized thus far. Our approach models activation patterns of thresholded nodes in (deep) artificial neural networks as stochastic processes. We focus solely on activation frequency, leveraging neuroscience techniques used for real neuron spike trains. During a classification task, we extract spiking activity and use an arrival process following the Poisson distribution. We examine observed data from various artificial neural networks in image recognition tasks, fitting the proposed model's assumptions. Through this, we derive parameters describing activation patterns in each network. Our analysis covers randomly initialized, generalizing, and memorizing networks, revealing consistent differences across architectures and training sets. Calculating Mean Firing Rate, Mean Fano Factor, and Variances, we find stable indicators of memorization during learning, providing valuable insights into network behavior. The proposed model shows promise in describing activation patterns and could serve as a general framework for future investigations. It has potential applications in theoretical simulations, pruning, and transfer learning.

cs.LG↗

arXiv2023

ContainerGym: A Real-World Reinforcement Learning Benchmark for Resource Allocation

We present ContainerGym, a benchmark for reinforcement learning inspired by a real-world industrial resource allocation task. The proposed benchmark encodes a range of challenges commonly encountered in real-world sequential decision making problems, such as uncertainty. It can be configured to instantiate problems of varying degrees of difficulty, e.g., in terms of variable dimensionality. Our benchmark differs from other reinforcement learning benchmarks, including the ones aiming to encode real-world difficulties, in that it is directly derived from a real-world industrial problem, which underwent minimal simplification and streamlining. It is sufficiently versatile to evaluate reinforcement learning algorithms on any real-world problem that fits our resource allocation framework. We provide results of standard baseline methods. Going beyond the usual training reward curves, our results and the statistical tools used to interpret them allow to highlight interesting limitations of well-known deep reinforcement learning algorithms, namely PPO, TRPO and DQN.

cs.LG↗

arXiv2022

Invariance to Quantile Selection in Distributional Continuous Control

In recent years distributional reinforcement learning has produced many state of the art results. Increasingly sample efficient Distributional algorithms for the discrete action domain have been developed over time that vary primarily in the way they parameterize their approximations of value distributions, and how they quantify the differences between those distributions. In this work we transfer three of the most well-known and successful of those algorithms (QR-DQN, IQN and FQF) to the continuous action domain by extending two powerful actor-critic algorithms (TD3 and SAC) with distributional critics. We investigate whether the relative performance of the methods for the discrete action space translates to the continuous case. To that end we compare them empirically on the pybullet implementations of a set of continuous control tasks. Our results indicate qualitative invariance regarding the number and placement of distributional atoms in the deterministic, continuous action setting.

cs.LG↗

arXiv2022

Recipe for Fast Large-scale SVM Training: Polishing, Parallelism, and more RAM!

Support vector machines (SVMs) are a standard method in the machine learning toolbox, in particular for tabular data. Non-linear kernel SVMs often deliver highly accurate predictors, however, at the cost of long training times. That problem is aggravated by the exponential growth of data volumes over time. It was tackled in the past mainly by two types of techniques: approximate solvers, and parallel GPU implementations. In this work, we combine both approaches to design an extremely fast dual SVM solver. We fully exploit the capabilities of modern compute servers: many-core architectures, multiple high-end GPUs, and large random access memory. On such a machine, we train a large-margin classifier on the ImageNet data set in 24 minutes.

cs.LG↗

arXiv2022

The (1+1)-ES Reliably Overcomes Saddle Points

It is known that step size adaptive evolution strategies (ES) do not converge (prematurely) to regular points of continuously differentiable objective functions. Among critical points, convergence to minima is desired, and convergence to maxima is easy to exclude. However, surprisingly little is known on whether ES can get stuck at a saddle point. In this work we establish that even the simple (1+1)-ES reliably overcomes most saddle points under quite mild regularity conditions. Our analysis is based on drift with tail bounds. It is non-standard in that we do not even aim to estimate hitting times based on drift. Rather, in our case it suffices to show that the relevant time is finite with full probability.

cs.NE↗

arXiv2022

ConTraNet: A single end-to-end hybrid network for EEG-based and EMG-based human machine interfaces

Objective: Electroencephalography (EEG) and electromyography (EMG) are two non-invasive bio-signals, which are widely used in human machine interface (HMI) technologies (EEG-HMI and EMG-HMI paradigm) for the rehabilitation of physically disabled people. Successful decoding of EEG and EMG signals into respective control command is a pivotal step in the rehabilitation process. Recently, several Convolutional neural networks (CNNs) based architectures are proposed that directly map the raw time-series signal into decision space and the process of meaningful features extraction and classification are performed simultaneously. However, these networks are tailored to the learn the expected characteristics of the given bio-signal and are limited to single paradigm. In this work, we addressed the question that can we build a single architecture which is able to learn distinct features from different HMI paradigms and still successfully classify them. Approach: In this work, we introduce a single hybrid model called ConTraNet, which is based on CNN and Transformer architectures that is equally useful for EEG-HMI and EMG-HMI paradigms. ConTraNet uses CNN block to introduce inductive bias in the model and learn local dependencies, whereas the Transformer block uses the self-attention mechanism to learn the long-range dependencies in the signal, which are crucial for the classification of EEG and EMG signals. Main results: We evaluated and compared the ConTraNet with state-of-the-art methods on three publicly available datasets which belong to EEG-HMI and EMG-HMI paradigms. ConTraNet outperformed its counterparts in all the different category tasks (2-class, 3-class, 4-class, and 10-class decoding tasks). Significance: The results suggest that ConTraNet is robust to learn distinct features from different HMI paradigms and generalizes well as compared to the current state of the art algorithms.

q-bio.NC↗

arXiv2022

From Motion to Muscle

Voluntary human motion is the product of muscle activity that results from upstream motion planning of the motor cortical areas. We show that muscle activity can be artificially generated based on motion features such as position, velocity, and acceleration. For this purpose, we specifically develop an approach based on a recurrent neural network trained in a supervised learning session; additional neural network architectures are considered and evaluated. The performance is evaluated by a new score called the zero-line score. The latter adaptively rescales the loss function of the generated signal for all channels by comparing the overall range of muscle activity and thus dynamically evaluates similarities between both signals. The model achieves a remarkable precision for previously trained motion while new motions that were not trained before still have high accuracy. Further, these models are trained on multiple subjects and thus are able to generalize across individuals. In addition, we distinguish between a general model that has been trained on several subjects, a subject-specific model, and a specific pre-trained model that uses the general model as a basis and is adapted to a specific subject afterward. The subject-specific generation of muscle activity can be further exploited to improve the rehabilitation of neuromuscular diseases with myoelectric prostheses and functional electric stimulation.

cs.LG↗

arXiv2022

Global Linear Convergence of Evolution Strategies on More Than Smooth Strongly Convex Functions

Evolution strategies (ESs) are zeroth-order stochastic black-box optimization heuristics invariant to monotonic transformations of the objective function. They evolve a multivariate normal distribution, from which candidate solutions are generated. Among different variants, CMA-ES is nowadays recognized as one of the state-of-the-art zeroth-order optimizers for difficult problems. Albeit ample empirical evidence that ESs with a step-size control mechanism converge linearly, theoretical guarantees of linear convergence of ESs have been established only on limited classes of functions. In particular, theoretical results on convex functions are missing, where zeroth-order and also first-order optimization methods are often analyzed. In this paper, we establish almost sure linear convergence and a bound on the expected hitting time of an \new{ES family, namely the $(1+1)_κ$-ES, which includes the (1+1)-ES with (generalized) one-fifth success rule} and an abstract covariance matrix adaptation with bounded condition number, on a broad class of functions. The analysis holds for monotonic transformations of positively homogeneous functions and of quadratically bounded functions, the latter of which particularly includes monotonic transformation of strongly convex functions with Lipschitz continuous gradient. As far as the authors know, this is the first work that proves linear convergence of ES on such a broad class of functions.

math.NA↗

arXiv2021

Deep Transfer-Learning for patient specific model re-calibration: Application to sEMG-Classification

Accurate decoding of surface electromyography (sEMG) is pivotal for muscle-to-machine-interfaces (MMI) and their application for e.g. rehabilitation therapy. sEMG signals have high inter-subject variability, due to various factors, including skin thickness, body fat percentage, and electrode placement. Therefore, obtaining high generalization quality of a trained sEMG decoder is quite challenging. Usually, machine learning based sEMG decoders are either trained on subject-specific data, or at least recalibrated for each user, individually. Even though, deep learning algorithms produced several state of the art results for sEMG decoding,however, due to the limited amount of availability of sEMG data, the deep learning models are prone to overfitting. Recently, transfer learning for domain adaptation improved generalization quality with reduced training time on various machine learning tasks. In this study, we investigate the effectiveness of transfer learning using weight initialization for recalibration of two different pretrained deep learning models on a new subjects data, and compare their performance to subject-specific models. To the best of our knowledge, this is the first study that thoroughly investigated weight-initialization based transfer learning for sEMG classification and compared transfer learning to subject-specific modeling. We tested our models on three publicly available databases under various settings. On average over all settings, our transfer learning approach improves 5~\%-points on the pretrained models without fine-tuning and 12~\%-points on the subject-specific models, while being trained on average 22~\% fewer epochs. Our results indicate that transfer learning enables faster training on fewer samples than user-specific models, and improves the performance of pretrained models as long as enough data is available.

cs.LG↗

arXiv2021

Anchored-STFT and GNAA: An extension of STFT in conjunction with an adversarial data augmentation technique for the decoding of neural signals

Brain-computer interfaces (BCIs) enable communication between humans and machines by translating brain activity into control commands. Electroencephalography (EEG) signals are one of the most used brain signals in non-invasive BCI applications but are often contaminated with noise. Therefore, it is possible that meaningful patterns for classifying EEG signals are deeply hidden. State-of-the-art deep-learning algorithms are successful in learning hidden, meaningful patterns. However, the quality and the quantity of the presented inputs is pivotal. Here, we propose a novel feature extraction method called anchored Short Time Fourier Transform (anchored-STFT), which is an advanced version of STFT, as it minimizes the trade-off between temporal and spectral resolution presented by STFT. In addition, we propose a novel augmentation method, called gradient norm adversarial augmentation (GNAA). GNAA is not only an augmentation method but is also used to harness adversarial inputs in EEG data, which not only improves the classification accuracy but also enhances the robustness of the classifier. In addition, we also propose a new CNN architecture, namely Skip-Net, for the classification of EEG signals. The proposed pipeline outperforms all state-of-the-art methods and yields an average classification accuracy of 90.7 % and 89.54 % on BCI competition II dataset III and BCI competition IV dataset 2b, respectively.

q-bio.QM↗

arXiv2021

Non-local Optimization: Imposing Structure on Optimization Problems by Relaxation

In stochastic optimization, particularly in evolutionary computation and reinforcement learning, the optimization of a function $f: Ω\to \mathbb{R}$ is often addressed through optimizing a so-called relaxation $θ\in Θ\mapsto \mathbb{E}_θ(f)$ of $f$, where $Θ$ resembles the parameters of a family of probability measures on $Ω$. We investigate the structure of such relaxations by means of measure theory and Fourier analysis, enabling us to shed light on the success of many associated stochastic optimization methods. The main structural traits we derive and that allow fast and reliable optimization of relaxations are the consistency of optimal values of $f$, Lipschitzness of gradients, and convexity. We emphasize settings where $f$ itself is not differentiable or convex, e.g., in the presence of (stochastic) disturbance.

math.OC↗

arXiv2021

Convergence Analysis of the Hessian Estimation Evolution Strategy

The class of algorithms called Hessian Estimation Evolution Strategies (HE-ESs) update the covariance matrix of their sampling distribution by directly estimating the curvature of the objective function. The approach is practically efficient, as attested by respectable performance on the BBOB testbed, even on rather irregular functions. In this paper we formally prove two strong guarantees for the (1+4)-HE-ES, a minimal elitist member of the family: stability of the covariance matrix update, and as a consequence, linear convergence on all convex quadratic problems at a rate that is independent of the problem instance.

math.OC↗

arXiv2021

Latent Representation Prediction Networks

Deeply-learned planning methods are often based on learning representations that are optimized for unrelated tasks. For example, they might be trained on reconstructing the environment. These representations are then combined with predictor functions for simulating rollouts to navigate the environment. We find this principle of learning representations unsatisfying and propose to learn them such that they are directly optimized for the task at hand: to be maximally predictable for the predictor function. This results in representations that are by design optimal for the downstream task of planning, where the learned predictor function is used as a forward model. To this end, we propose a new way of jointly learning this representation along with the prediction function, a system we dub Latent Representation Prediction Network (LARP). The prediction function is used as a forward model for search on a graph in a viewpoint-matching task and the representation learned to maximize predictability is found to outperform a pre-trained representation. Our approach is shown to be more sample-efficient than standard reinforcement learning methods and our learned representation transfers successfully to dissimilar objects.

cs.LG↗

arXiv2020

Global Convergence of the (1+1) Evolution Strategy

We establish global convergence of the (1+1) evolution strategy, i.e., convergence to a critical point independent of the initial state. More precisely, we show the existence of a critical limit point, using a suitable extension of the notion of a critical point to measurable functions. At its core, the analysis is based on a novel progress guarantee for elitist, rank-based evolutionary algorithms. By applying it to the (1+1) evolution strategy we are able to provide an accurate characterization of whether global convergence is guaranteed with full probability, or whether premature convergence is possible. We illustrate our results on a number of example applications ranging from smooth (non-convex) cases over different types of saddle points and ridge functions to discontinuous and extremely rugged problems.

cs.NE↗

arXiv2020

Data Augmentation and Clustering for Vehicle Make/Model Classification

Vehicle shape information is very important in Intelligent Traffic Systems (ITS). In this paper we present a way to exploit a training data set of vehicles released in different years and captured under different perspectives. Also the efficacy of clustering to enhance the make/model classification is presented. Both steps led to improved classification results and a greater robustness. Deeper convolutional neural network based on ResNet architecture has been designed for the training of the vehicle make/model classification. The unequal class distribution of training data produces an a priori probability. Its elimination, obtained by removing of the bias and through hard normalization of the centroids in the classification layer, improves the classification results. A developed application has been used to test the vehicle re-identification on video data manually based on make/model and color classification. This work was partially funded under the grant.

cs.CV↗

arXiv2020

Methods of the Vehicle Re-identification

Most of researchers use the vehicle re-identification based on classification. This always requires an update with the new vehicle models in the market. In this paper, two types of vehicle re-identification will be presented. First, the standard method, which needs an image from the search vehicle. VRIC and VehicleID data set are suitable for training this module. It will be explained in detail how to improve the performance of this method using a trained network, which is designed for the classification. The second method takes as input a representative image of the search vehicle with similar make/model, released year and colour. It is very useful when an image from the search vehicle is not available. It produces as output a shape and a colour features. This could be used by the matching across a database to re-identify vehicles, which look similar to the search vehicle. To get a robust module for the re-identification, a fine-grained classification has been trained, which its class consists of four elements: the make of a vehicle refers to the vehicle's manufacturer, e.g. Mercedes-Benz, the model of a vehicle refers to type of model within that manufacturer's portfolio, e.g. C Class, the year refers to the iteration of the model, which may receive progressive alterations and upgrades by its manufacturer and the perspective of the vehicle. Thus, all four elements describe the vehicle at increasing degree of specificity. The aim of the vehicle shape classification is to classify the combination of these four elements. The colour classification has been separately trained. The results of vehicle re-identification will be shown. Using a developed tool, the re-identification of vehicles on video images and on controlled data set will be demonstrated. This work was partially funded under the grant.

cs.CV↗

arXiv2020

The Hessian Estimation Evolution Strategy

We present a novel black box optimization algorithm called Hessian Estimation Evolution Strategy. The algorithm updates the covariance matrix of its sampling distribution by directly estimating the curvature of the objective function. This algorithm design is targeted at twice continuously differentiable problems. For this, we extend the cumulative step-size adaptation algorithm of the CMA-ES to mirrored sampling. We demonstrate that our approach to covariance matrix adaptation is efficient by evaluation it on the BBOB/COCO testbed. We also show that the algorithm is surprisingly robust when its core assumption of a twice continuously differentiable objective function is violated. The approach yields a new evolution strategy with competitive performance, and at the same time it also offers an interesting alternative to the usual covariance matrix update mechanism.

cs.LG↗

arXiv2020

Analyzing Reinforcement Learning Benchmarks with Random Weight Guessing

We propose a novel method for analyzing and visualizing the complexity of standard reinforcement learning (RL) benchmarks based on score distributions. A large number of policy networks are generated by randomly guessing their parameters, and then evaluated on the benchmark task; the study of their aggregated results provide insights into the benchmark complexity. Our method guarantees objectivity of evaluation by sidestepping learning altogether: the policy network parameters are generated using Random Weight Guessing (RWG), making our method agnostic to (i) the classic RL setup, (ii) any learning algorithm, and (iii) hyperparameter tuning. We show that this approach isolates the environment complexity, highlights specific types of challenges, and provides a proper foundation for the statistical analysis of the task's difficulty. We test our approach on a variety of classic control benchmarks from the OpenAI Gym, where we show that small untrained networks can provide a robust baseline for a variety of tasks. The networks generated often show good performance even without gradual learning, incidentally highlighting the triviality of a few popular benchmarks.

cs.LG↗