Large Language Models (LLMs) have become the starting point for many Natural Language Processing tasks, offering strong performance without task-specific training. However, they are computationally expensive and struggle with numerical reasoning. To tackle these issues, we propose GLiNER-Reg, a regression model adapted from GLiNER2 that derives continuous predictions by calculating the expected value over a set of reference prompts. Our results demonstrate that GLiNER-Reg outperforms the original GLiNER2 baseline in both fine-tuning and zero-shot regression scenarios, offering a highly efficient regression model alternative to LLMs.

Contemporary grand strategists, political scientists, and economists have formulated various theoretical models to analyse historical events and forecast or predict future scenarios for strategic decision-makers. However, many of these models remain predominantly qualitative and narrative-led. While they provide some valuable principles and domain knowledge, they often lack empirical quantification, which limits their applications for making precise decisions over time. Traditional frameworks usually struggle to address issues such as interconnectivity, dynamics, nonlinearity, feedback, emergencies, co-evolution, unpredictability, uncertainty, and ambiguity. To address this gap, we propose a hierarchical framework underpinned by seven decision layers, which can be quantified by Theory-Informed Machine Learning (TIML) methods. This framework enables us to manage various grand strategies or strategic challenges. We argue that a grand strategy is an abstract pattern of human intelligence that emerges from multiple decision layers below and is driven by emotional rewards from above. It transcends the way of balancing means with the end and context. We consider that a grand strategy is a type of strange attractor, which is a deterministic chaos. Determinism implies that our time, resources, capability, and cultural background are bounded. Chaos entails unpredictable long-term consequences. We aim to create a new computational model that can craft a robust grand strategy driven by TIML based on chaos and complexity theories.

LLM-based Multi-Agent Systems (MAS) achieve strong performance on complex reasoning tasks by coordinating multiple agents, but at the cost of substantial token consumption. Recent work on automatic topology design, ARG-Designer, has reframed this problem as autoregressive graph generation. However, its training objective provides no explicit incentive for the model to generate sparse and efficient topologies. We address this limitation by introducing a Reward-Guided Autoregressive Graph Generation (RGA-Designer) inspired by Reinforcement Learning from Human Feedback (RLHF). We train a reward model that jointly captures task correctness and structural compactness, and then fine-tune the pretrained graph generator using the reward model as feedback. Our method preserves task accuracy at the level of ARG-Designer while reducing token consumption by an average of 20.5

We propose LQ-SGD (Low-Rank Quantized Stochastic Gradient Descent), an efficient communication gradient compression algorithm designed for distributed training. LQ-SGD further develops on the basis of PowerSGD by incorporating the low-rank approximation and log-quantization techniques, which drastically reduce the communication overhead, while still ensuring the convergence speed of training and model accuracy. In addition, LQ-SGD and other compression-based methods show stronger resistance to gradient inversion than traditional SGD, providing a more robust and efficient optimization path for distributed learning systems.

Distributed learning (DL) uses multiple nodes to accelerate training, enabling efficient optimization of large-scale models. Stochastic Gradient Descent (SGD), a key optimization algorithm, plays a central role in this process. However, communication bottlenecks often limit scalability and efficiency, leading to increasing adoption of compressed SGD techniques to alleviate these challenges. Despite addressing communication overheads, compressed SGD introduces trustworthiness concerns, as gradient exchanges among nodes are vulnerable to attacks like gradient inversion (GradInv) and membership inference attacks (MIA). The trustworthiness of compressed SGD remains unexplored, leaving important questions about its reliability unanswered. In this paper, we provide a trustworthiness evaluation of compressed versus uncompressed SGD. Specifically, we conducted empirical studies using GradInv attacks, revealing that compressed SGD demonstrates significantly higher resistance to privacy leakage compared to uncompressed SGD. In addition, our findings suggest that MIA may not be a reliable metric for assessing privacy risks in distributed learning.

Transformer-based models for time series forecasting (TSF) have attracted significant attention in recent years due to their effectiveness and versatility. However, these models often require extensive hyperparameter optimization (HPO) to achieve the best possible performance, and a unified pipeline for HPO in transformer-based TSF remains lacking. In this paper, we present one such pipeline and conduct extensive experiments on several state-of-the-art (SOTA) transformer-based TSF models. These experiments are conducted on standard benchmark datasets to evaluate and compare the performance of different models, generating practical insights and examples. Our pipeline is generalizable beyond transformer-based architectures and can be applied to other SOTA models, such as Mamba and TimeMixer, as demonstrated in our experiments. The goal of this work is to provide valuable guidance to both industry practitioners and academic researchers in efficiently identifying optimal hyperparameters suited to their specific domain applications. The code and complete experimental results are available on GitHub.

Ensuring data trustworthiness within individual edge nodes while facilitating collaborative data processing poses a critical challenge in edge computing systems (ECS), particularly in resource-constrained scenarios such as autonomous systems sensor networks, industrial IoT, and smart cities. This paper presents a lightweight, fully distributed k-means clustering algorithm specifically adapted for edge environments, leveraging a distributed averaging approach with additive secret sharing, a secure multiparty computation technique, during the cluster center update phase to ensure the accuracy and trustworthiness of data across nodes.

Many machine learning (ML) models can make predictions regarding credit default swaps (CDS) for the telecommunication (telco) service sector. However, some ML algorithms can only offer a black-box model. It is crucial to explain the prediction result for strategic decisions. We compare various the state of arts, including deep learning (transformers), gradient boost machine (GBM), and Xgboost, plus different explainable tools: Variable Importance (VI) Partial Dependent Plots (PDP), Local Individual Conditional Expectation (LIME), Interpretable Model-agnostic Explanations (ICE), and Shapley values for the prediction model. Moreover, we also conducted a hyperparameter search of the prediction model by leveraging high-performance computing (HPC). Our experiment results show that the Xgboost provides the best solution with fewer constraints. We aim to find an optimal solution for strategic CDS investment decisions.

Alongside the continuous process of improving AI performance through the development of more sophisticated models, researchers have also focused their attention to the emerging concept of data-centric AI, which emphasizes the important role of data in a systematic machine learning training process. Nonetheless, the development of models has also continued apace. One result of this progress is the development of the Transformer Architecture, which possesses a high level of capability in multiple domains such as Natural Language Processing (NLP), Computer Vision (CV) and Time Series Forecasting (TSF). Its performance is, however, heavily dependent on input data preprocessing and output data evaluation, justifying a data-centric approach to future research. We argue that data-centric AI is essential for training AI models, particularly for transformer-based TSF models efficiently. However, there is a gap regarding the integration of transformer-based TSF and data-centric AI. This survey aims to pin down this gap via the extensive literature review based on the proposed taxonomy. We review the previous research works from a data-centric AI perspective and we intend to lay the foundation work for the future development of transformer-based architecture and data-centric AI.

We often use "explainable" Artificial Intelligence (XAI)" and "interpretable AI (IAI)" interchangeably when we apply various XAI tools for a given dataset to explain the reasons that underpin machine learning (ML) outputs. However, these notions can sometimes be confusing because interpretation often has a subjective connotation, while explanations lean towards objective facts. We argue that XAI is a subset of IAI. The concept of IAI is beyond the sphere of a dataset. It includes the domain of a mindset. At the core of this ambiguity is the duality of reasons, in which we can reason either outwards or inwards. When directed outwards, we want the reasons to make sense through the laws of nature. When turned inwards, we want the reasons to be happy, guided by the laws of the heart. While XAI and IAI share reason as the common notion for the goal of transparency, clarity, fairness, reliability, and accountability in the context of ethical AI and trustworthy AI (TAI), their differences lie in that XAI emphasizes the post-hoc analysis of a dataset, and IAI requires a priori mindset of abstraction. This hypothesis can be proved by empirical experiments based on an open dataset and harnessed by High-Performance Computing (HPC). The demarcation of XAI and IAI is indispensable because it would be impossible to determine regulatory policies for many AI applications, especially in healthcare, human resources, banking, and finance. We aim to clarify these notions and lay the foundation of XAI, IAI, EAI, and TAI for many practitioners and policymakers in future AI applications and research.

Many machine learning (ML) models can make predictions regarding credit default swaps (CDS) for the telecommunication (telco) service sector. However, some ML algorithms can only offer a black-box model. It is crucial to explain the prediction result for strategic decisions. We compare various the state of arts, including deep learning (transformers), gradient boost machine (GBM), and Xgboost, plus different explainable tools: Variable Importance (VI) Partial Dependent Plots (PDP), Local Individual Conditional Expectation (LIME), Interpretable Model-agnostic Explanations (ICE), and Shapley values for the prediction model. Moreover, we also conducted a hyperparameter search of the prediction model by leveraging high-performance computing (HPC). Our experiment results show that the Xgboost provides the best solution with fewer constraints. We aim to find an optimal solution for strategic CDS investment decisions.

In the domain of multivariate forecasting, transformer models stand out as powerful apparatus, displaying exceptional capabilities in handling messy datasets from real-world contexts. However, the inherent complexity of these datasets, characterized by numerous variables and lengthy temporal sequences, poses challenges, including increased noise and extended model runtime. This paper focuses on reducing redundant information to elevate forecasting accuracy while optimizing runtime efficiency. We propose a novel transformer forecasting framework enhanced by Principal Component Analysis (PCA) to tackle this challenge. The framework is evaluated by five state-of-the-art (SOTA) models and four diverse real-world datasets. Our experimental results demonstrate the framework's ability to minimize prediction errors across all models and datasets while significantly reducing runtime. From the model perspective, one of the PCA-enhanced models: PCA+Crossformer, reduces mean square errors (MSE) by 33.3 perspective, the framework delivers 14.3 Electricity datasets, as well as 4.8 Traffic datasets. This study aims to advance various SOTA models and enhance transformer-based time series forecasting for intricate data.

When engaging in strategic decision-making, we are frequently confronted with overwhelming information and data. The situation can be further complicated when certain pieces of evidence contradict each other or become paradoxical. The primary challenge is how to determine which information can be trusted when we adopt Artificial Intelligence (AI) systems for decision-making. This issue is known as deciding what to decide or Trustworthy AI. However, the AI system itself is often considered an opaque black box. We propose a new approach to address this issue by introducing a novel framework of Trustworthy AI (TAI) encompassing three crucial components of AI: representation space, loss function, and optimizer. Each component is loosely coupled with four TAI properties. Altogether, the framework consists of twelve TAI properties. We aim to use this framework to conduct the TAI experiments by quantitive and qualitative research methods to satisfy TAI properties for the decision-making context. The framework allows us to formulate an optimal prediction model trained by the given dataset for applying the strategic investment decision of credit default swaps (CDS) in the technology sector. Finally, we provide our view of the future direction of TAI research

When engaging in strategic decision-making, we are frequently confronted with overwhelming information and data. The situation can be further complicated when certain pieces of evidence contradict each other or become paradoxical. The primary challenge is how to determine which information can be trustedwhenwe adopt Artificial Intelligence (AI) systems for decision-making. This issue is known as "deciding what to decide" or Trustworthy AI. However, the AI system itself is often considered an opaque "black box". We propose a new approach to address this issue by introducing a novel framework of Trustworthy AI (TAI) encompassing three crucial components of AI: representation space, loss function, and optimizer. Each component is loosely coupled with four TAI properties. Altogether, the framework consists of twelve TAI properties. We aim to use this framework to conduct the TAI experiments by quantitive and qualitative research methods to satisfy TAI properties for the decision-making context. The framework allows us to formulate an optimal prediction model trained by the given dataset for applying the strategic investment decision of credit default swaps (CDS) in the technology sector. Finally, we provide our view of the future direction of TAI research.

Artificial intelligence (AI) advancements allow machines to achieve human-like intelligence. Problem-solving and decision-making are two mental abilities to measure human intelligence. Building a generalized representational model for various inputs and outputs is essential to obtaining such humanlike capabilities. Many scholars tried to articulate different models from different perspectives. However, there is a gap in establishing an overall AI-oriented hierarchical framework. This study proposes a novel model known as the emerged AI protocol that consists of seven abstractive layers capable of providing a meaningful solution for a given problem. In contrast to previous hierarchies, we argue that this unique model is conceptually evident, logically consistent, theoretically compelling, and practically adaptable. We aim to create a generalized model that can be implemented by various machine learning (ML) algorithms for problem-solving and decision-making.

When making strategic decisions, we are often confronted with overwhelming information to process. The situation can be further complicated when some pieces of evidence are contradicted each other or paradoxical. The challenge then becomes how to determine which information is useful and which ones should be eliminated. This process is known as meta-decision. Likewise, when it comes to using Artificial Intelligence (AI) systems for strategic decision-making, placing trust in the AI itself becomes a meta-decision, given that many AI systems are viewed as opaque "black boxes" that process large amounts of data. Trusting an opaque system involves deciding on the level of Trustworthy AI (TAI). We propose a new approach to address this issue by introducing a novel taxonomy or framework of TAI, which encompasses three crucial domains: articulate, authentic, and basic for different levels of trust. To underpin these domains, we create ten dimensions to measure trust: explainability/transparency, fairness/diversity, generalizability, privacy, data governance, safety/robustness, accountability, reproducibility, reliability, and sustainability. We aim to use this taxonomy to conduct a comprehensive survey and explore different TAI approaches from a strategic decision-making perspective.

Strategic Decision-Making is always challenging because it is inherently uncertain, ambiguous, risky, and complex. By contrast to tactical and operational decisions, strategic decisions are decisive, pivotal, and often irreversible, which may result in long-term and significant consequences. A strategic decision-making process usually involves many aspects of inquiry, including sensory perception, deliberative thinking, inquiry-based analysis, meta-learning, and constant interaction with the external world. Many unknowns, unpredictabilities, and environmental constraints will shape every aspect of a strategic decision. Traditionally, this task often relies on intuition, reflective thinking, visionary insights, approximate estimates, and practical wisdom. With recent advances in artificial intelligence/machine learning (AI/ML) technologies, we can leverage AI/ML to support strategic decision-making. However, there is still a substantial gap from an AI perspective due to inadequate models, despite the tremendous progress made. We argue that creating a comprehensive taxonomy of decision frames as a representation space is essential for AI because it could offer surprising insights beyond anyone's imaginary boundary today. Strategic decision-making is the art of possibility. This study develops a systematic taxonomy of decision-making frames that consists of six bases, 18 categorical, and 54 elementary frames. We formulate the model using the inquiry method based on Bloom's taxonomy approach. We aim to lay out the computational foundation that is possible to capture a comprehensive landscape view of a strategic problem. Compared with many traditional models, this novel taxonomy covers irrational, non-rational and rational frames capable of dealing with certainty, uncertainty, complexity, ambiguity, chaos, and ignorance.

Cloud service providers (CSP) and cloud consumers often need to forecast the cloud price to optimize their business strategy. However, pricing of cloud services is a challenging task due to its services complexity and dynamic nature of the ever-changing environment. Moreover, the cloud pricing based on consumers’ willingness to pay (W2P) becomes even more challenging due to the subjectiveness of consumers’ experiences and implicit values of some non-marketable features, such as burstable CPU, dedicated server, and cloud data center global footprints. Unfortunately, many existing pricing models often cannot support value-based pricing. In this paper, we propose a novel solution based on value-based pricing, which does not only consider how much does the service cost (or intrinsic values) to a CSP but also how much a customer is willing to pay (or extrinsic values) for the service. We demonstrate that the cloud extrinsic values would not only become one of the competitive advantages for CSPs to lead the cloud market but also increase the profit margin. Our approach is often referred to as a hedonic pricing model. We show that our model can capture the value of non-marketable features. This value is about 43.4 percent on average above the baseline, which is often ignored by many traditional cloud pricing models. We also show that Average Annual Growth Rate (AAGR) of Amazon Web Services’ (AWS) is about -20.0 percent per annum between 2008 and 2017, ceteris paribus. In comparison with Moore's law (-50 percent per annum), it is at a far slower pace. We argue this value is Moore's law equivalent in the cloud. The primary goal of this research is to provide a less biased pricing model for cloud decision makers to develop their optimizing investment strategy.

Modeling a utility function for cloud business customers is one of the critical challenges facing many cloud service providers (CSPs) for their pricing strategy. It concerns how to measure various subjective experiences of the business customers and how to translate their cloud service experiences into a quantifiable unit, which can be determined by a utility function that reflects cloud resource consumption. The aim of this modeling process is to set up a pricing foundation so that CSPs can target a broader range of customers from various market segments and identify the optimal price point of their various pricing models. Previous studies have either focused on simple theoretical proof or drifted the meaning of utility between demand and supply or proposed a solution based on a uniform cloud market assumption. This paper proposes a novel and practical solution to define multiple utility functions based on a scenario of six cloud market segments, which are analyzed by three analytic approaches that are known as Markov chains analysis, queueing theory, and risk assessment. The entire pricing strategy emphasizes value co-creation between CSP and cloud business customers. In comparison with other methods, such as calibrated, price-quality, resource-based, simple linear, and capacity-aware, this method provides both internal and external rationalities for CSP to capture the subjective value of cloud business customers. Consequently, our experiment results show that this modeling method can increase a profit margin by 51% and decrease a unit cost by 22% for a CSP.