Abstract--Understanding the robustness of deep learning models for multivariate long-term time series forecasting (ML TSF) remains challenging, as evaluations typically rely on real-world datasets with unknown noise properties. We propose a simulation-based evaluation framework that generates parameterizable synthetic datasets, where each dataset instance corresponds to a different configuration of signal components, noise types, signal-to-noise ratios, and frequency characteristics. These configurable components aim to model real-world multivariate time series data without the ambiguity of unknown noise. This framework enables fine-grained, systematic evaluation of M-L TSF models under controlled and diverse scenarios. Our analysis reveals that all models degrade severely when lookback windows cannot capture complete periods of seasonal patters in the data. S-Mamba and Autoformer perform best on sawtooth patterns, while R-Linear and iTransformer favor sinusoidal signals. White and Brownian noise universally degrade performance with lower signal-to-noise ratio while S-Mamba shows specific trend-noise and iTransformer shows seasonal-noise vulnerability. Further spectral analysis shows that S-Mamba and iTransformer achieve superior frequency reconstruction. This controlled approach, based on our synthetic and principle-driven testbed, offers deeper insights into model-specific strengths and limitations through the aggregation of MSE scores and provides concrete guidance for model selection based on signal characteristics and noise conditions. IME series forecasting plays a crucial role across diverse fields such as energy systems [1]-[3], meteorology [4], [5], traffic flow modeling [6], [7] or the modeling of sensor networks [8], [9]. Reliable forecasts support proactive decision-making, effective risk management, and efficient planning. As high-resolution temporal data becomes increasingly available, the need for robust and scalable forecasting models has grown more important than ever. A time series represents data points ordered in time and can be categorized as either univariate, when consisting of a single variable, or multivariate, when involving multiple interdependent variables [10].

Modern anthropomorphic upper limb bioprostheses are typically controlled by electromyographic (EMG) biosignals using a pattern recognition scheme. Unfortunately, there are many factors originating from the human source of objects to be classified and from the human-prosthesis interface that make it difficult to obtain an acceptable classification quality. One of these factors is the high susceptibility of biosignals to contamination, which can considerably reduce the quality of classification of a recognition system. In the paper, the authors propose a new recognition system intended for EMG based control of the hand prosthesis with detection of contaminated biosignals in order to mitigate the adverse effect of contaminations. The system consists of two ensembles: the set of one-class classifiers (OCC) to assess the degree of contamination of individual channels and the ensemble of K-nearest neighbours (KNN) classifier to recognise the patient's intent. For all recognition systems, an original, coherent fuzzy model was developed, which allows the use of a uniform soft (fuzzy) decision scheme throughout the recognition process. The experimental evaluation was conducted using real biosignals from a public repository. The goal was to provide an experimental comparative analysis of the parameters and procedures of the developed method on which the quality of the recognition system depends. The proposed fuzzy recognition system was also compared with similar systems described in the literature.

The accessibility of long-duration recorders, adapted to sometimes demanding field conditions, has enabled the deployment of extensive animal population monitoring campaigns through ecoacoustics. The effectiveness of automatic signal detection methods, increasingly based on neural approaches, is frequently evaluated solely through machine learning metrics, while acoustic analysis of performance remains rare. As part of the acoustic monitoring of Rock Ptarmigan populations, we propose here a simple method for acoustic analysis of the detection system's performance. The proposed measure is based on relating the signal-to-noise ratio of synthetic signals to their probability of detection. We show how this measure provides information about the system and allows optimisation of its training. We also show how it enables modelling of the detection distance, thus offering the possibility of evaluating its dynamics according to the sound environment and accessing an estimation of the spatial density of calls.

Best subset selection in linear regression is well known to be nonconvex and computationally challenging to solve, as the number of possible subsets grows rapidly with increasing dimensionality of the problem. As a result, finding the global optimal solution via an exact optimization method for a problem with dimensions of 1000s may take an impractical amount of CPU time. This suggests the importance of finding suboptimal procedures that can provide good approximate solutions using much less computational effort than exact methods. In this work, we introduce a new procedure and compare it with other popular suboptimal algorithms to solve the best subset selection problem. Extensive computational experiments using synthetic and real data have been performed. The results provide insights into the performance of these methods in different data settings. The new procedure is observed to be a competitive suboptimal algorithm for solving the best subset selection problem for high-dimensional data.

Adam is the go-to optimizer for training modern machine learning models, but it requires additional memory to maintain the moving averages of the gradients and their squares. While various low-memory optimizers have been proposed that sometimes match the performance of Adam, their lack of reliability has left Adam as the default choice. In this work, we apply a simple layer-wise Signal-to-Noise Ratio (SNR) analysis to quantify when second-moment tensors can be effectively replaced by their means across different dimensions. Our SNR analysis reveals how architecture, training hyperparameters, and dataset properties impact compressibility along Adam's trajectory, naturally leading to $\textit{SlimAdam}$, a memory-efficient Adam variant. $\textit{SlimAdam}$ compresses the second moments along dimensions with high SNR when feasible, and leaves when compression would be detrimental. Through experiments across a diverse set of architectures and training scenarios, we show that $\textit{SlimAdam}$ matches Adam's performance and stability while saving up to $98\%$ of total second moments. Code for $\textit{SlimAdam}$ is available at https://github.com/dayal-kalra/low-memory-adam.

Myopotential pattern recognition to decode the intent of the user is the most advanced approach to controlling a powered bioprosthesis. Unfortunately, many factors make this a difficult problem and achieving acceptable recognition quality in real-word conditions is a serious challenge. The aim of the paper is to develop a recognition system that will mitigate factors related to multimodality and multichannel recording of biosignals and their high susceptibility to contamination. The proposed method involves the use of two co-operating multiclassifier systems. The first system is composed of one-class classifiers related to individual electromyographic (EMG) and mechanomyographic (MMG) biosignal recording channels, and its task is to recognise contaminated channels. The role of the second system is to recognise the class of movement resulting from the patient's intention. The ensemble system consists of base classifiers using the representation (extracted features) of biosignals from different channels. The system uses a dynamic selection mechanism, eliminating those base classifiers that are associated with biosignal channels that are recognised by the one-class ensemble system as being contaminated. Experimental studies were conducted using signals from an able-bodied person with simulation of amputation. The results obtained allow us to reject the null hypothesis that the application of the dual ensemble foes not lead to improved classification quality.

This work proposes a novel channel estimator based on diffusion models (DMs), one of the currently top-rated generative models. Contrary to related works utilizing generative priors, a lightweight convolutional neural network (CNN) with positional embedding of the signal-to-noise ratio (SNR) information is designed by learning the channel distribution in the sparse angular domain. Combined with an estimation strategy that avoids stochastic resampling and truncates reverse diffusion steps that account for lower SNR than the given pilot observation, the resulting DM estimator has both low complexity and memory overhead. Numerical results exhibit better performance than state-of-the-art channel estimators utilizing generative priors.

In this work we propose RELDEC, a novel approach for sequential decoding of moderate length low-density parity-check (LDPC) codes. The main idea behind RELDEC is that an optimized decoding policy is subsequently obtained via reinforcement learning based on a Markov decision process (MDP). In contrast to our previous work, where an agent learns to schedule only a single check node (CN) within a group (cluster) of CNs per iteration, in this work we train the agent to schedule all CNs in a cluster, and all clusters in every iteration. That is, in each learning step of RELDEC an agent learns to schedule CN clusters sequentially depending on a reward associated with the outcome of scheduling a particular cluster. We also modify the state space representation of the MDP, enabling RELDEC to be suitable for larger block length LDPC codes than those studied in our previous work. Furthermore, to address decoding under varying channel conditions, we propose agile meta-RELDEC (AM-RELDEC) that employs meta-reinforcement learning. The proposed RELDEC scheme significantly outperforms standard flooding and random sequential decoding for a variety of LDPC codes, including codes designed for 5G new radio.

State-of-the-art performance for many edge applications is achieved by deep neural networks (DNNs). Often, these DNNs are location-and time-sensitive, and must be delivered over a wireless channel rapidly and efficiently. In this paper, we introduce AirNet, a family of novel training and transmission methods that allow DNNs to be efficiently delivered over wireless channels under stringent transmit power and latency constraints. This corresponds to a new class of joint source-channel coding problems, aimed at delivering DNNs with the goal of maximizing their accuracy at the receiver, rather than recovering them with high fidelity. In AirNet, we propose the direct mapping of the DNN parameters to transmitted channel symbols, while the network is trained to meet the channel constraints, and exhibit robustness against channel noise. AirNet achieves higher accuracy compared to separation-based alternatives. We further improve the performance of AirNet by pruning the network below the available bandwidth, and expanding it for improved robustness. We also benefit from unequal error protection by selectively expanding important layers of the network. Finally, we develop an approach, which simultaneously trains a spectrum of DNNs, each targeting a different channel condition, resolving the impractical memory requirements of training distinct networks for different channel conditions. The results in this paper were presented in part at the 2022 IEEE International Symposium on Information Theory (ISIT) [1]. Developments within the area of DL have been made possible mainly thanks to the rapid growth of the computational power and memory available for both the researchers and the potential users of various DL-based algorithms. This resulted in the development of increasingly complex deep neural networks (DNNs) with millions and even billions of parameters trained on massive datasets, achieving impressive accuracy and performance in a wide variety of applications. On the other hand, the memory required to store a single modern DNN model can easily go from a few megabytes up to hundreds of gigabytes.

EEG signals are complex and low-frequency signals. Therefore, they are easily influenced by external factors. EEG artifact removal is crucial in neuroscience because artifacts have a significant impact on the results of EEG analysis. The removal of ocular artifacts is the most challenging among these artifacts. In this study, a novel ocular artifact removal method is presented by developing bidirectional long-short term memory (BiLSTM)-based deep learning (DL) models. We created a benchmarking dataset to train and test proposed DL models by combining the EEGdenoiseNet and DEAP datasets. We also augmented the data by contaminating ground-truth clean EEG signals with EOG at various SNR levels. The BiLSTM network is then fed to features extracted from augmented signals using highly-localized time-frequency (TF) coefficients obtained by wavelet synchrosqueezed transform (WSST). We also compare the WSST-based DL model results with traditional TF analysis (TFA) methods namely short-time Fourier transformation (STFT) and continuous wavelet transform (CWT) as well as augmented raw signals. The best average MSE value of 0.3066 was obtained by the first time-proposed BiLSTM-based WSST-Net model. Our results demonstrated the WSST-Net model significantly improves artifact removal performance compared to traditional TF and raw signal methods. Also, the proposed EOG removal approach reveals that it outperforms many conventional and DL-based ocular artifact removal methods in the literature.