Zohaib Khan, Mustafa Dogan, Ifeoma Okoh, Pouya Sadeghi, Siddhartha Shrestha, Sergius Justus Chesami Nyah, Mahmoud O. Mokhiamar, Michael J Ryan, Tarek Naous. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

Although generative AI is being deployed into classrooms with promises of aiding teachers, educators caution that these tools can have unintended pedagogical repercussions, including cultural misrepresentation and bias. These concerns are heightened in low-resource language and Indigenous education settings, where AI systems frequently underperform. We investigate these challenges in Hawai‘i, where public schools operate under a statewide mandate to integrate Hawaiian language and culture into education. Through four co-design workshops with 22 public school educators, we surfaced concerns about using generative AI in educational settings, particularly around cultural misrepresentation, and corresponding designs for auditing tools that address these issues. We find that educators envision tools grounded in specific Hawaiian cultural values and practices, such as tracing the genealogy of knowledge in source materials. Building on these insights, we conceptualize AI auditing as a community-oriented process rather than the work of isolated individuals, and discuss implications for designing auditing tools.

Voice assistants, such as Siri and Google Assistant, typically model audio and text separately, resulting in lost speech information and increased complexity. Recent efforts to address this with end-to-end Speech Large Language Models (speech-in, text-out) trained with supervised finetuning (SFT) have led to models “forgetting” capabilities from text-only LLMs. Our work proposes an alternative paradigm for training Speech LLMs without instruction data, using the response of a text-only LLM to transcripts as self-supervision. Importantly, this process can be performed without annotated responses. We show that our Distilled Voice Assistant (DiVA) generalizes to Spoken Question Answering, Classification, and Translation. Furthermore, DiVA better matches user preferences, achieving a 72% win rate compared with state-of-the-art models like Qwen 2 Audio, despite using >100x less training compute.

Minzhi Li, William Held, Michael J. Ryan, Kunat Pipatanakul, Potsawee Manakul, Hao Zhu, Diyi Yang. Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2025.

Michael J. Ryan, Omar Shaikh, Aditri Bhagirath, Daniel Frees, William Held, Diyi Yang. Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2025.

As the reach of large language models (LMs) expands globally, their ability to cater to diverse cultural contexts becomes crucial.Despite advancements in multilingual capabilities, models are not designed with appropriate cultural nuances.In this paper, we show that multilingual and Arabic monolingual LMs exhibit bias towards entities associated with Western culture.We introduce CAMeL, a novel resource of 628 naturally-occurring prompts and 20,368 entities spanning eight types that contrast Arab and Western cultures.CAMeL provides a foundation for measuring cultural biases in LMs through both extrinsic and intrinsic evaluations.Using CAMeL, we examine the cross-cultural performance in Arabic of 16 different LMs on tasks such as story generation, NER, and sentiment analysis, where we find concerning cases of stereotyping and cultural unfairness.We further test their text-infilling performance, revealing the incapability of appropriate adaptation to Arab cultural contexts.Finally, we analyze 6 Arabic pre-training corpora and find that commonly used sources such as Wikipedia may not be best suited to build culturally aware LMs, if used as they are without adjustment.We will make CAMeL publicly available at: https://github.com/tareknaous/camel

Before being deployed for user-facing applications, developers align Large Language Models (LLMs) to user preferences through a variety of procedures, such as Reinforcement Learning From Human Feedback (RLHF) and Direct Preference Optimization (DPO).Current evaluations of these procedures focus on benchmarks of instruction following, reasoning, and truthfulness.However, human preferences are not universal, and aligning to specific preference sets may have unintended effects.We explore how alignment impacts performance along three axes of global representation: English dialects, multilingualism, and opinions from and about countries worldwide.Our results show that current alignment procedures create disparities between English dialects and global opinions.We find alignment improves capabilities in several languages.We conclude by discussing design decisions that led to these unintended impacts and recommendations for more equitable preference tuning.We make our code and data publicly available on Github 1 .

We present a comprehensive evaluation of large language models for multilingual readability assessment. Existing evaluation resources lack domain and language diversity, limiting the ability for cross-domain and cross-lingual analyses. This paper introduces ReadMe++, a multilingual multi-domain dataset with human annotations of 9757 sentences in Arabic, English, French, Hindi, and Russian, collected from 112 different data sources. This benchmark will encourage research on developing robust multilingual readability assessment methods. Using ReadMe++, we benchmark multilingual and monolingual language models in the supervised, unsupervised, and few-shot prompting settings. The domain and language diversity in ReadMe++ enable us to test more effective few-shot prompting, and identify shortcomings in state-of-the-art unsupervised methods. Our experiments also reveal exciting results of superior domain generalization and enhanced cross-lingual transfer capabilities by models trained on ReadMe++. We will make our data publicly available and release a python package tool for multilingual sentence readability prediction using our trained models at: https://github.com/tareknaous/readme.

Krista Opsahl-Ong, Michael J Ryan, Josh Purtell, David Broman, Christopher Potts, Matei Zaharia, Omar Khattab. Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing. 2024.

Recent advancements in high-quality, largescale English resources have pushed the frontier of English Automatic Text Simplification (ATS) research.However, less work has been done on multilingual text simplification due to the lack of a diverse evaluation benchmark that covers complex-simple sentence pairs in many languages.This paper introduces the MULTI-SIM benchmark, a collection of 27 resources in 12 distinct languages containing over 1.7 million complex-simple sentence pairs.This benchmark will encourage research in developing more effective multilingual text simplification models and evaluation metrics.Our experiments using MULTISIM with pre-trained multilingual language models reveal exciting performance improvements from multilingual training in non-English settings.We observe strong performance from Russian in zero-shot crosslingual transfer to low-resource languages.We further show that few-shot prompting with BLOOM-176b achieves comparable quality to reference simplifications outperforming finetuned models in most languages.We validate these findings through human evaluation.

With continued improvements in wireless sensing technology, the notion of the Internet of Things (IoT) has been widely adopted and has become pervasive owing to its broad applications in scenarios such as ambient assisted living, smart healthcare, and smart homes. In that regard, human activity recognition (HAR) is a vital element of intelligent systems to undertake persistent surveillance of human behavior. Due to the omnipresent impact of smartphones in each person’s life, smartphone inertial sensors are used as a case study for this research. Most of the conventional approaches regard HAR as a time-series classification problem; yet, the accuracy of recognition degrades for heterogeneous sensors. In this article, we investigate encoding sensory heterogeneous HAR (HHAR) data into three-channel image representation (i.e., RGB), hence treat the HHAR task as an image classification problem. Since present convolutional network models are computationally heavy when deployed in the IoT environment, we propose a lightweight model image encoded HHAR, called multiscale image-encoded HHAR (MS-IE-HHAR). The model employs a hierarchical multiscale extraction (HME) module followed by an improved spatialwise and channelwise attention (ISCA) module to form the main architecture of the model. The HME module is formed by a group of residually connected shuffle group convolutions (SG-Conv) to extract and learn image representations from different receptive fields while reducing the number of network parameters. The ISCA module combines a lightweight spatialwise attention (SwA) block and an improved channelwise attention (CwA) module to enable the network to pay instructive attention to spatial correlations as well as channel interdependency information. Finally, two widely available HHAR public data sets (i.e., HHAR UCI, and MHEALTH) were used to evaluate the performance of the proposed models with accuracy over 98% and 99%, respectively, demonstrating the model superiority for modeling HAR from heterogeneous data sources.

Abstract Feature Selection (FS) is an important preprocessing step that is involved in machine learning and data mining tasks for preparing data (especially high-dimensional data) by eliminating irrelevant and redundant features, thus reducing the potential curse of dimensionality of a given large dataset. Consequently, FS is arguably a combinatorial NP-hard problem in which the computational time increases exponentially with an increase in problem complexity. To tackle such a problem type, meta-heuristic techniques have been opted by an increasing number of scholars. Herein, a novel meta-heuristic algorithm, called Sparrow Search Algorithm (SSA), is presented. The SSA still performs poorly on exploratory behavior and exploration-exploitation trade-off because it does not duly stimulate the search within feasible regions, and the exploitation process suffers noticeable stagnation. Therefore, we improve SSA by adopting: i) a strategy for Random Re-positioning of Roaming Agents (3RA); and ii) a novel Local Search Algorithm (LSA), which are algorithmically incorporated into the original SSA structure. To the FS problem, SSA is improved and cloned as a binary variant, namely, the improved Binary SSA (iBSSA), which would strive to select the optimal or near-optimal features from a given dataset while keeping the classification accuracy maximized. For binary conversion, the iBSSA was primarily validated against nine common S-shaped and V-shaped Transfer Functions (TFs), thus producing nine iBSSA variants. To verify the robustness of these variants, three well-known classification techniques, includingk-Nearest Neighbor (k-NN), Support Vector Machine (SVM), and Random Forest (RF) were adopted as fitness evaluators with the proposed iBSSA approach and many other competing algorithms, on 18 multifaceted, multi-scale benchmark datasets from the University of California Irvine (UCI) data repository. Then, the overall best-performing iBSSA variant for each of the three classifiers was compared with binary variants of 12 different well-known meta-heuristic algorithms, including the original SSA (BSSA), Artificial Bee Colony (BABC), Particle Swarm Optimization (BPSO), Bat Algorithm (BBA), Grey Wolf Optimization (BGWO), Whale Optimization Algorithm (BWOA), Grasshopper Optimization Algorithm (BGOA) SailFish Optimizer (BSFO), Harris Hawks Optimization (BHHO), Bird Swarm Algorithm (BBSA), Atom Search Optimization (BASO), and Henry Gas Solubility Optimization (BHGSO). Based on a Wilcoxon’s non-parametric statistical test ( $$\alpha =0.05$$ α=0.05 ), the superiority of iBSSA with the three classifiers was very evident against counterparts across the vast majority of the selected datasets, achieving a feature size reduction of up to 92% along with up to 100% classification accuracy on some of those datasets.

Processing data from Internet of Things (IoT) applications at the cloud centers has known limitations relating to latency, task scheduling, and load balancing. Hence, there have been a shift towards adopting fog computing as a complementary paradigm to cloud systems. In this article, we first propose a multi-objective task-scheduling optimization problem that minimizes both the makespans and total costs in a fog-cloud environment. Then, we suggest an optimization model based on a Discrete Non-dominated Sorting Genetic Algorithm II (DNSGA-II) to deal with the discrete multi-objective task-scheduling problem and to automatically allocate tasks that should be executed either on fog or cloud nodes. The NSGA-II algorithm is adapted to discretize crossover and mutation evolutionary operators, rather than using continuous operators that require high computational resources and not able to allocate proper computing nodes. In our model, the communications between the fog and cloud tiers are formulated as a multi-objective function to optimize the execution of tasks. The proposed model allocates computing resources that would effectively run on either the fog or cloud nodes. Moreover, it efficiently organizes the distribution of workloads through various computing resources at the fog. Several experiments are conducted to determine the performance of the proposed model compared with a continuous NSGA-II (CNSGA-II) algorithm and four peer mechanisms. The outcomes demonstrate that the model is capable of achieving dynamic task scheduling with minimizing the total execution times (i.e., makespans) and costs in fog-cloud environments.

Real-time forecasting of the financial time-series data is challenging for many machine learning (ML) algorithms. First, many ML models operate offline, where they need a batch of data, which may not be available during training. Besides, due to a fixed architecture of the majority of the offline-based ML models, they suffer to deal with the uncertain nature of financial time-series data. In contrast, online learning mode evolving-structured ML models could be promising for financial time-series forecasting. For real-time deployment of such models, low memory demand is a must. Besides, the model’s explainability plays a crucial role in forecasting financial time-series. Considering all the requirements, a rule-based autonomous neuro-fuzzy learning algorithm called the parsimonious learning machine (PALM) is proposed here to forecast time-varying stock indices. To provide efficient automation of the proposed algorithm by maintaining the model explainability in terms of limited number linguistic IF-THEN rules, two popular multiobjective evolutionary algorithms (MEAs), such as a real-coded genetic algorithm (GA) and a self-adaptive differential evolution (DE) algorithm are utilized here. In addition, fuzzy type-2 variants of PALMs’ are considered here due to better uncertainty handling capacity than their type-1 counterparts. To evaluate the proposed algorithm’s performance, the closing stock price of fifteen (15) different stock market indices are predicted here. From the results, it is observed that the MEA-based PALMs are performing better than the state-of-the-art benchmark online ML models and providing a rule-based explainable model to the end-user.

Increases in demand for a greater variety of products help companies gain more shares of growing competitive markets but, in contrast, lead to an increase in production processes and, therefore, higher costs and longer lead times. Although several techniques for platform formations and assembly lines have been introduced to enable more varieties of goods to be produced, this also makes a system more complex and less cost-efficient. This paper proposes a differential evolution (DE) approach that incorporates a new heuristic method, improved solution representation and enhanced crossover and mutation operators for solving the modular-based product family design problem in a reconfigurable manufacturing system. The heuristic is applied to repair some solutions in the initial population by replacing eligible components with packages to provide near-optimal solutions in the initial stage and enable DE to find the optimal solution quickly. The proposed crossover is designed to further use the repaired solutions to produce new individuals with better qualities. Finally, a case study of a kettle family is conducted to validate this heuristic method, with the experimental results showing that it saves 57.5% of the purchasing costs of components and, on average, 41.35% of setup costs compared with those of median-joining phylogenetic network- and non-platform-based heuristics. Moreover, the proposed DE achieves improved performances with average errors of 63.34% and 38.52% from those of the standard versions of DE and a genetic algorithm, respectively, in terms of the total production costs of producing the same variants.

In project planning, traditionally, the project managers first schedule the project activities, and then plan for materials ordering and supplier selection. This disintegrated approach lacks in planning co-ordination and causes a loss in expected profit for the organization. In this study, a concurrent model is proposed for resource constraint project scheduling with materials ordering and suppliers selection problems. The proposed model aims to maximize overall net present value for the organization considering ordering cost, procurement cost, materials holding cost, and the project’s deadline penalty cost. The resource constraint project scheduling with materials ordering and suppliers selection is an NP-hard problem. Thus, a memetic algorithm hybridizing the genetic algorithm with a forward-backward improvement based local search is proposed to solve the proposed model. The proposed algorithm is tested on self-generated 120 instances varied from 30 to 60 activity projects with 4 to 8 suppliers. Experimental results show that the proposed genetic algorithm-based memetic algorithm approach generates better solutions than the standalone genetic algorithm. The concurrent project scheduling with materials ordering and supplier selection approach and solution methods have a significant implication for the managers to complete the project in an economic and timely manner.

Modern information technology, such as the internet of things (IoT) provides a real-time experience into how a system is performing and has been used in diversified areas spanning from machines, supply chain, and logistics to smart cities. IoT captures the changes in surrounding environments based on collections of distributed sensors and then sends the data to a fog computing (FC) layer for analysis and subsequent response. The speed of decision in such a process relies on there being minimal delay, which requires efficient distribution of tasks among the fog nodes. Since the utility of FC relies on the efficiency of this task scheduling task, improvements are always being sought in the speed of response. Here, we suggest an improved elitism genetic algorithm (IEGA) for overcoming the task scheduling problem for FC to enhance the quality of services to users of IoT devices. The improvements offered by IEGA stem from two main phases: first, the mutation rate and crossover rate are manipulated to help the algorithms in exploring most of the combinations that may form the near-optimal permutation; and a second phase mutates a number of solutions based on a certain probability to avoid becoming trapped in local minima and to find a better solution. IEGA is compared with five recent robust optimization algorithms in addition to EGA in terms of makespan, flow time, fitness function, carbon dioxide emission rate, and energy consumption. IEGA is shown to be superior to all other algorithms in all respects.

Human activity recognition (HAR) has been regarded as an indispensable part of many smart home systems and smart healthcare applications. Specifically, HAR is of great importance in the Internet of Healthcare Things (IoHT), owing to the rapid proliferation of Internet of Things (IoT) technologies embedded in various smart appliances and wearable devices (such as smartphones and smartwatches) that have a pervasive impact on an individual's life. The inertial sensors of smartphones generate massive amounts of multidimensional time-series data, which can be exploited effectively for HAR purposes. Unlike traditional approaches, deep learning techniques are the most suitable choice for such multivariate streams. In this study, we introduce a supervised dual-channel model that comprises long short-term memory (LSTM), followed by an attention mechanism for the temporal fusion of inertial sensor data concurrent with a convolutional residual network for the spatial fusion of sensor data. We also introduce an adaptive channel-squeezing operation to fine-tune convolutional a neural network feature extraction capability by exploiting multichannel dependency. Finally, two widely available and public HAR data sets are used in experiments to evaluate the performance of our model. The results demonstrate that our proposed approach can overcome state-of-the-art methods.

The rapid growth of the Internet of Things (IoT) technologies has generated a huge amount of traffic that can be exploited for detecting intrusions through IoT networks. Despite the great effort made in annotating IoT traffic records, the number of labeled records is still very small, increasing the difficulty in recognizing attacks and intrusions. This study introduces a semi-supervised deep learning approach for intrusion detection (SS-Deep-ID), in which we propose a multiscale residual temporal convolutional (MS-Res) module to finetune the network capability in learning spatiotemporal representations. An improved traffic attention (TA) mechanism is introduced to estimate the importance score that helps the model to concentrate on important information during learning. Furthermore, a hierarchical semi-supervised training method is introduced which takes into account the sequential characteristics of the IoT traffic data during training. The proposed SS-Deep-ID is easily integrated into a fog-enabled IoT network to offer efficient real-time intrusion detection. Finally, empirical evaluations on two recent data sets (CIC-IDS2017 and CIC-IDS2018) demonstrate that SS-Deep-ID improves the efficiency of intrusion detection and increases the robustness of performance while maintaining computational efficiency.

Although intelligent load forecasting is essential for optimal energy management (EM) in smart cities, there is a lack of current research exploring EM in well-regulated Internet-of-Things (IoT) networks. This article develops a new deep learning (DL) model for efficient forecasting of short-term energy consumption while maintaining effective communication between energy providers and users. The proposed Energy-Net stack comprises multiple stacked spatiotemporal modules, where each module consists of a temporal transformer (TT) submodule and a spatial transformer (ST) submodule. The TT models the temporal relationships in load data; and the ST submodule extracts hidden spatial information by integrating convolutional layers and includes an improved self-attention mechanism. The experimental evaluation on IHPEC and independent system operator New England (ISO-NE) data set demonstrates the superiority of Energy-Net over recent cutting-edge DL models with root mean-square error (RMSE) of 0.354 and 0.535, respectively. The computational complexity of Energy-Net is appropriate for dependable resource-constrained IoT devices (i.e., fog nodes or edge nodes) linked to a joint IoT-cloud server that interacts with connected smart grids to handle EM tasks.

The extensive propagation of industrial Internet of Things (IIoT) technologies has encouraged intruders to initiate a variety of attacks that need to be identified to maintain the security of end-user data and the safety of services offered by service providers. Deep learning (DL), especially recurrent approaches, has been applied successfully to the analysis of IIoT forensics but their key challenge of recurrent DL models is that they struggle with long traffic sequences and cannot be parallelized. Multihead attention (MHA) tried to address this shortfall but failed to capture the local representation of IIoT traffic sequences. In this article, we propose a forensics-based DL model (called Deep-IFS) to identify intrusions in IIoT traffic. The model learns local representations using local gated recurrent unit (LocalGRU), and introduces an MHA layer to capture and learn global representation (i.e., long-range dependencies). A residual connection between layers is designed to prevent information loss. Another challenge facing the current IIoT forensics frameworks is their limited scalability, limiting performance in handling Big IIoT traffic data produced by IIoT devices. This challenge is addressed by deploying and training the proposed Deep-IFS in a fog computing environment. The intrusion identification becomes scalable by distributing the computation and the IIoT traffic data across worker fog nodes for training the model. The master fog node is responsible for sharing training parameters and aggregating worker node output. The aggregated classification output is subsequently passed to the cloud platform for mitigating attacks. Empirical results on the Bot-IIoT dataset demonstrate that the developed distributed Deep-IFS can effectively handle Big IIoT traffic data compared with the present centralized DL-based forensics techniques. Further, the results validate the robustness of the proposed Deep-IFS across various evaluation measures.

In recent years, several multi-method and multi-operator-based algorithms have been proposed for solving optimization problems. Generally, their performance is better than other algorithms that based on a single operator and/or algorithm. However, they do not perform consistently well over all the problems tested in the literature. In this paper, we propose an improved optimization algorithm that uses the benefits of multiple differential evolution operators, with more emphasis placed on the best-performing operator. The performance of the proposed algorithm is tested by solving 10 problems with 5, 10, 15 and 20 dimensions taken from CEC2020 competition on single objective bound constrained optimization, with its results outperforming both single operator-based and different state-of-the-art algorithms.

Recently, many deferential evolution-based algorithms have been developed to solve constrained optimization problems. The performance of these methods outperforms the performance of single operator and/or algorithm-based ones. However, they do not perform consistently for all the problems tested in the literature. Also, the process of using the appropriate selection of algorithms and operators may be time-consuming since their designs are undertaken mainly through trial and error. In this paper, we propose an improved optimization algorithm that uses the benefits of multiple deferential evolution operators, with the best one is emphasized based on the quality and diversity of the population. The performance of the proposed algorithm is tested by solving 57 real-world constrained problems with different dimensions, number of equality and equality constraints, with its results showing a high success rate and that it outperformed different state-of-the-art algorithms.

In this paper, we model and solve a strategic problem of fleet renewal to meet future operational needs under uncertain conditions. The fleet renewal problem focuses on mainly strategic decisions involving from fleet size, fleet mix and timing of replacement, yet it is essential to consider a significant amount of detail regarding short-term decisions to prevent inferior or infeasible strategies. In this direction, we develop a hybrid simulation model by combining system dynamics (SD) and discrete event simulation (DES) approaches. The standalone use of this model enables the decision maker to analyze the effects of both short- and long-term decisions on availability by simulating the processes that the fleet undertakes through its life-cycle from asset acquisition to retirement. Nevertheless, the simulation neither suggests nor seeks the best renewal strategy(ies). To alleviate this difficulty, we propose a simulation-based optimization that uses a genetic algorithm (GA) to effectively search a very large set of feasible fleet renewal strategies and uses the developed hybrid simulation model to evaluate candidate strategies found by GA. To provide a decision context where the approach has been developed and applied, we use a naval fleet renewal application. The extensive numerical experiments show that the proposed approach not only finds good and robust renewal strategies but also identify critical resources that influence the fleet’s availability. Finally, the robustness of optimized strategies under uncertainty is tested by sensitivity analysis, and mappings between implemented strategies and the fleet performance are constructed by scenario discovery analysis to provide insights for decision makers.

This study investigates the makespan minimization problem that arises when compiling a general class of quantum algorithms into near-term quantum hardware. The problem is referred to as quantum circuit compilation problem (QCCP). Two new heuristics are proposed for solving the problem. Firstly, a rollout based sequential decision making approach is investigated. This heuristic decides on which operation to schedule next based on the makespan projection given by a guiding priority rule. Secondly, a stochastic version of rollout heuristic is proposed which iteratively switches between rollout and a simple priority rule to explore the search space. The two proposed heuristics are able to show improvement in makespan across a range of instance sizes and characteristics.

The resource constraint project scheduling problem (RCPSP) is one of the complex combinatorial optimization problems. Several approaches have been proposed over the decades to solve RCPSPs optimally within reasonable computational times. In this paper, a genetic algorithm based memetic algorithm (MA) is proposed to solve RCPSP with makespan minimization as the objective. The proposed algorithm incorporates Nawaz, Enscore, and Ham (NEH) heuristic-based initialization with carefully designed crossover and local search techniques. The performance of the proposed approach is evaluated by solving 1560 problem instances from the popular project scheduling library (PSPLIB) and the results for different datasets have been compared with selected state-of-the-art algorithms. The comparison demonstrates the satisfactory performance of the proposed approach. Extensive experiments reveal that the proposed algorithm is very simple to implement and highly effective when compared to state-of-the-art methods.

The field of many-objective optimization has grown out of infancy and a number of contemporary algorithms can deliver well converged and diverse sets of solutions close to the Pareto optimal front. Concurrently, the studies in cognitive science have highlighted the pitfalls of imprecise decision-making in presence of a large number of alternatives. Thus, for effective decision-making, it is important to devise methods to identify a handful (7 ± 2) of solutions from a potentially large set of tradeoff solutions. Existing measures such as reflex/bend angle, expected marginal utility (EMU), maximum convex bulge/distance from hyperplane, hypervolume contribution, and local curvature are inadequate for the purpose as: 1) they may not create complete ordering of the solutions; 2) they cannot deal with large number of objectives and/or solutions; and 3) they typically do not provide any insight on the nature of selected solutions (internal, peripheral, and extremal). In this letter, we introduce a scheme to identify solutions of interest based on recursive use of the EMU measure. The nature of the solutions (internal or peripheral) is then characterized using reference directions generated via systematic sampling and the top K solutions with the largest relative EMU measure are presented to the decision maker. The performance of the approach is illustrated using a number of benchmarks and engineering problems. In our opinion, the development of such methods is necessary to bridge the gap between theoretical development and real-world adoption of many-objective optimization algorithms.

While an increasing number of satellite images are collected over a regular period in order to provide regular spatiotemporal information on land-use and land-cover changes, there are very few compression schemes in remotely sensed imagery that use historical data as a reference. Just as individual images can be compressed for separate transmission by taking into account their inherent spatial and spectral redundancies, the temporal redundancy between images of the same scene can also be exploited for sequential transmission. In this letter, we propose a nonlinear elastic method based on the general relationship to predict adaptively the current image from a previous reference image without any loss of information. The main feature of the developed method is to find the best prediction for each pixel brightness value individually using its own conditional probabilities to the previous image, instead of applying a single linear or nonlinear model. A codebook is generated to record the nonlinear point-to-point relationship. This temporal lossless compression is incorporated with spatial- and spectral-domain predictions, and the performances are compared with those of the JPEG2000 standard. The experimental results show an improved performance by more than 5%.

ABSTRACT For many applications, it is desirable for a node to be able to identify the neighbor nodes currently within range. However, the identification procedures for terrestrial communication networks (TCNs) are inefficient when implemented in long‐delay networks (LDNs) such as underwater acoustic networks or satellite networks, where the propagation time of a packet cannot be neglected. Here, we propose a time‐efficient and power‐efficient procedure, which is insensitive to propagation delay, for identifying neighbors in an LDN by optimizing the network offered load using either unslotted or slotted protocols. The procedure is adapted to the scenario when node cardinality and distribution are either known or unknown in advance. We achieve improvements as high as 80% in time and 45% in power consumption for our proposed approach compared with approaches developed previously for TCN. Furthermore, we analyze our proposed approach with the inclusion of the capture effect and packet receive time variations. We also provide a closed form formula for finding the optimum guard time and a procedure to estimate the packet receive time variations. Copyright © 2012 John Wiley & Sons, Ltd.

Images compressed by current lossy compression techniques suffer from distortion that is uniformly distributed spatially and spectrally. We demonstrate that classification accuracy changes as a function of pixel class and spectral location (spectral bands) of the distortion and that the uniform distribution of distortion is therefore not an optimal approach to controlling distortion within an image. We show that a superior approach involves locating areas within the image whose classification accuracies are relatively insensitive to distortion and limiting the application of distortion to these areas. By following this approach, we show that substantial levels of compression-induced distortion can be tolerated without a significant reduction in subsequent classification accuracy. Hyperspectral images are prime candidates for data compression due to their inherent size. Lossy compression algorithms are attractive as they typically provide the most impressive levels of compression (compression ratio) but result in some distortion of the original image. The acceptability of this distortion depends on the end use of the data set that, in the case of hyperspectral images, invariably involves the use of computer-based tools to process the images into a set of classes representing the ground coverage or conditions present in the data. Compression-induced distortion tends to reduce the accuracy associated with the classification process, but the relationship between distortion and classification accuracy varies across different classes of data within the same image and is not particularly predictable. We demonstrate an alternative to the uniform application of distortion during compression that aims to locate spectral and spatial areas within an image where sensitivity to distortion is likely to be reduced. We then restrict the application of the compression-induced distortion to these areas of low sensitivity and show that the subsequent classification accuracies are superior to uniformly distorted images.

Mean-normalized vector quantization (M-NVQ) has been demonstrated to be the preferred technique for lossless compression of hyperspectral data. In this paper, a jointly optimized spatial M-NVQ/spectral DCT technique is shown to produce compression ratios significantly better than those obtained by the optimized spatial M-NVQ technique alone.

Mean-normalised vector quantization (M-NVQ) has been demonstrated to be the preferred vector quantization technique for application to the lossless compression of hyperspectral data. This work optimises M-NVQ parameters for application to lossy compression and slight improvement is shown to be gained by the implementation of spatial and spectral discrete cosine transform (DCT) techniques for coding of the M-NVQ residuals. Much more efficient compression is shown to be obtained by optimising the M-NVQ and DCT techniques simultaneously, rather than sequentially. Optimised spatial M-NVQ/spectral DCT is shown to produce compression ratios of between 1.5 and 2.5 times better than those obtained by the spatial M-NVQ technique alone. Compression ratios of up to 43:1 are achieved without significant loss in classification accuracy.

The structure of hyperspectral images reveals spectral responses that would seem ideal candidates for compression by vector quantization. This paper outlines the results of an investigation of lossless vector quantization of 224-band Airborne/Visible Infrared imaging Spectrometer (AVIRIS) images. Various vector formation techniques are identified and suitable quantization parameters are investigated. A new technique, mean-normalized vector quantization (M-NVQ), is proposed which produces compression performances approaching the theoretical minimum compressed image entropy of 5 bits/pixel. Images are compressed from original image entropies of between 8.28 and 10.89 bits/pixel to between 4.83 and 5.90 bits/pixel.