Recent language models use subwording mechanisms to handle Out-of-Vocabulary(OOV) words seen during test time and, their generation capacity is generally measured using perplexity, an intrinsic metric. It is known that increasing the subword granularity results in a decrease of perplexity value. However, the study of how subwording affects the understanding capacity of language models has been very few and only limited to a handful of languages. To reduce this gap we used 6 different tokenization schemes to pretrain relatively small language models in Nepali and used the representations learned to finetune on several downstream tasks. Although byte-level BPE algorithm has been used in recent models like GPT, RoBERTa we show that on average they are sub-optimal in comparison to algorithms such as SentencePiece in finetuning performances for Nepali. Additionally, similar recent studies have focused on the Bert-based language model. We, however, pretrain and finetune sequential transformer-based language models.

Land cover analysis using hyperspectral images (HSI) remains an open problem due to their low spatial resolution and complex spectral information. Recent studies are primarily dedicated to designing Transformer-based architectures for spatial-spectral long-range dependencies modeling, which is computationally expensive with quadratic complexity. Selective structured state space model (Mamba), which is efficient for modeling long-range dependencies with linear complexity, has recently shown promising progress. However, its potential in hyperspectral image processing that requires handling numerous spectral bands has not yet been explored. In this paper, we innovatively propose S$^2$Mamba, a spatial-spectral state space model for hyperspectral image classification, to excavate spatial-spectral contextual features, resulting in more efficient and accurate land cover analysis. In S$^2$Mamba, two selective structured state space models through different dimensions are designed for feature extraction, one for spatial, and the other for spectral, along with a spatial-spectral mixture gate for optimal fusion. More specifically, S$^2$Mamba first captures spatial contextual relations by interacting each pixel with its adjacent through a Patch Cross Scanning module and then explores semantic information from continuous spectral bands through a Bi-directional Spectral Scanning module. Considering the distinct expertise of the two attributes in homogenous and complicated texture scenes, we realize the Spatial-spectral Mixture Gate by a group of learnable matrices, allowing for the adaptive incorporation of representations learned across different dimensions. Extensive experiments conducted on HSI classification benchmarks demonstrate the superiority and prospect of S$^2$Mamba. The code will be available at: https://github.com/PURE-melo/S2Mamba.

Combustion instability in gas turbines and rocket engines, as one of the most challenging problems in combustion research, arises from the complex interactions among flames, which are also influenced by chemical reactions, heat and mass transfer, and acoustics. Identifying and understanding combustion instability is essential to ensure the safe and reliable operation of many combustion systems, where exploring and classifying the dynamical behaviors of complex flame systems is a core take. To facilitate fundamental studies, the present work concerns dynamical mode recognition of coupled flame oscillators made of flickering buoyant diffusion flames, which have gained increasing attention in recent years but are not sufficiently understood. The time series data of flame oscillators are generated by fully validated reacting flow simulations. Due to limitations of expertise-based models, a data-driven approach is adopted. In this study, a nonlinear dimensional reduction model of variational autoencoder (VAE) is used to project the simulation data onto a 2-dimensional latent space. Based on the phase trajectories in latent space, both supervised and unsupervised classifiers are proposed for datasets with well known labeling and without, respectively. For labeled datasets, we establish the Wasserstein-distance-based classifier (WDC) for mode recognition; for unlabeled datasets, we develop a novel unsupervised classifier (GMM-DTWC) combining dynamic time warping (DTW) and Gaussian mixture model (GMM). Through comparing with conventional approaches for dimensionality reduction and classification, the proposed supervised and unsupervised VAE-based approaches exhibit a prominent performance for distinguishing dynamical modes, implying their potential extension to dynamical mode recognition of complex combustion problems.

Fluid dynamics problems are characterized by being multidimensional and nonlinear, causing the experiments and numerical simulations being complex, time-consuming and monetarily expensive. In this sense, there is a need to find new ways to obtain data in a more economical manner. Thus, in this work we study the application of time series forecasting to fluid dynamics problems, where the aim is to predict the flow dynamics using only past information. We focus our study on models based on deep learning that do not require a high amount of data for training, as this is the problem we are trying to address. Specifically in this work we have tested three autoregressive models where two of them are fully based on deep learning and the other one is a hybrid model that combines modal decomposition with deep learning. We ask these models to generate $200$ time-ahead predictions of two datasets coming from a numerical simulation and experimental measurements, where the latter is characterized by being turbulent. We show how the hybrid model generates more reliable predictions in the experimental case, as it is physics-informed in the sense that the modal decomposition extracts the physics in a way that allows us to predict it.

Artificial intelligence is currently a dominant force in shaping various aspects of the world. Machine learning is a sub-field in artificial intelligence. Feature scaling is one of the data pre-processing techniques that improves the performance of machine learning algorithms. The traditional feature scaling techniques are unsupervised where they do not have influence of the dependent variable in the scaling process. In this paper, we have presented a novel feature scaling technique named DTization that employs decision tree and robust scaler for supervised feature scaling. The proposed method utilizes decision tree to measure the feature importance and based on the importance, different features get scaled differently with the robust scaler algorithm. The proposed method has been extensively evaluated on ten classification and regression datasets on various evaluation matrices and the results show a noteworthy performance improvement compared to the traditional feature scaling methods.

Logs are important in modern software development with runtime information. Log parsing is the first step in many log-based analyses, that involve extracting structured information from unstructured log data. Traditional log parsers face challenges in accurately parsing logs due to the diversity of log formats, which directly impacts the performance of downstream log-analysis tasks. In this paper, we explore the potential of using Large Language Models (LLMs) for log parsing and propose LLMParser, an LLM-based log parser based on generative LLMs and few-shot tuning. We leverage four LLMs, Flan-T5-small, Flan-T5-base, LLaMA-7B, and ChatGLM-6B in LLMParsers. Our evaluation of 16 open-source systems shows that LLMParser achieves statistically significantly higher parsing accuracy than state-of-the-art parsers (a 96% average parsing accuracy). We further conduct a comprehensive empirical analysis on the effect of training size, model size, and pre-training LLM on log parsing accuracy. We find that smaller LLMs may be more effective than more complex LLMs; for instance where Flan-T5-base achieves comparable results as LLaMA-7B with a shorter inference time. We also find that using LLMs pre-trained using logs from other systems does not always improve parsing accuracy. While using pre-trained Flan-T5-base shows an improvement in accuracy, pre-trained LLaMA results in a decrease (decrease by almost 55% in group accuracy). In short, our study provides empirical evidence for using LLMs for log parsing and highlights the limitations and future research direction of LLM-based log parsers.

Reservoir computing is a brain-inspired machine learning framework for processing temporal data by mapping inputs into high-dimensional spaces. Physical reservoir computers (PRCs) leverage native fading memory and nonlinearity in physical substrates, including atomic switches, photonics, volatile memristors, and, recently, memcapacitors, to achieve efficient high-dimensional mapping. Traditional PRCs often consist of homogeneous device arrays, which rely on input encoding methods and large stochastic device-to-device variations for increased nonlinearity and high-dimensional mapping. These approaches incur high pre-processing costs and restrict real-time deployment. Here, we introduce a novel heterogeneous memcapacitor-based PRC that exploits internal voltage offsets to enable both monotonic and non-monotonic input-state correlations crucial for efficient high-dimensional transformations. We demonstrate our approach's efficacy by predicting a second-order nonlinear dynamical system with an extremely low prediction error (0.00018). Additionally, we predict a chaotic H\'enon map, achieving a low normalized root mean square error (0.080). Unlike previous PRCs, such errors are achieved without input encoding methods, underscoring the power of distinct input-state correlations. Most importantly, we generalize our approach to other neuromorphic devices that lack inherent voltage offsets using externally applied offsets to realize various input-state correlations. Our approach and unprecedented performance are a major milestone towards high-performance full in-materia PRCs.

In the realm of time series forecasting (TSF), it is imperative for models to adeptly discern and distill hidden patterns within historical time series data to forecast future states. Transformer-based models exhibit formidable efficacy in TSF, primarily attributed to their advantage in apprehending these patterns. However, the quadratic complexity of the Transformer leads to low computational efficiency and high costs, which somewhat hinders the deployment of the TSF model in real-world scenarios. Recently, Mamba, a selective state space model, has gained traction due to its ability to process dependencies in sequences while maintaining near-linear complexity. For TSF tasks, these characteristics enable Mamba to comprehend hidden patterns as the Transformer and reduce computational overhead compared to the Transformer. Therefore, we propose a Mamba-based model named Simple-Mamba (S-Mamba) for TSF. Specifically, we tokenize the time points of each variate autonomously via a linear layer. A bidirectional Mamba layer is utilized to extract inter-variate correlations and a Feed-Forward Network is set to learn temporal dependencies. Finally, the generation of forecast outcomes through a linear mapping layer. Experiments on thirteen public datasets prove that S-Mamba maintains low computational overhead and achieves leading performance. Furthermore, we conduct extensive experiments to explore Mamba's potential in TSF tasks. Our code is available at https://github.com/wzhwzhwzh0921/S-D-Mamba.

Emerging data-driven approaches, such as deep reinforcement learning (DRL), aim at on-the-field learning of powertrain control policies that optimize fuel economy and other performance metrics. Indeed, they have shown great potential in this regard for individual vehicles on specific routes or drive cycles. However, for fleets of vehicles that must service a distribution of routes, DRL approaches struggle with learning stability issues that result in high variances and challenge their practical deployment. In this paper, we present a novel framework for shared learning among a fleet of vehicles through the use of a distilled group policy as the knowledge sharing mechanism for the policy learning computations at each vehicle. We detail the mathematical formulation that makes this possible. Several scenarios are considered to analyze the functionality, performance, and computational scalability of the framework with fleet size. Comparisons of the cumulative performance of fleets using our proposed shared learning approach with a baseline of individual learning agents and another state-of-the-art approach with a centralized learner show clear advantages to our approach. For example, we find a fleet average asymptotic improvement of 8.5 percent in fuel economy compared to the baseline while also improving on the metrics of acceleration error and shifting frequency for fleets serving a distribution of suburban routes. Furthermore, we include demonstrative results that show how the framework reduces variance within a fleet and also how it helps individual agents adapt better to new routes.

Characterizing and predicting the training performance of modern machine learning (ML) workloads on compute systems with compute and communication spread between CPUs, GPUs, and network devices is not only the key to optimization and planning but also a complex goal to achieve. The primary challenges include the complexity of synchronization and load balancing between CPUs and GPUs, the variance in input data distribution, and the use of different communication devices and topologies (e.g., NVLink, PCIe, network cards) that connect multiple compute devices, coupled with the desire for flexible training configurations. Built on top of our prior work for single-GPU platforms, we address these challenges and enable multi-GPU performance modeling by incorporating (1) data-distribution-aware performance models for embedding table lookup, and (2) data movement prediction of communication collectives, into our upgraded performance modeling pipeline equipped with inter-and intra-rank synchronization for ML workloads trained on multi-GPU platforms. Beyond accurately predicting the per-iteration training time of DLRM models with random configurations with a geomean error of 5.21% on two multi-GPU platforms, our prediction pipeline generalizes well to other types of ML workloads, such as Transformer-based NLP models with a geomean error of 3.00%. Moreover, even without actually running ML workloads like DLRMs on the hardware, it is capable of generating insights such as quickly selecting the fastest embedding table sharding configuration (with a success rate of 85%).