The Recursive Edge: A Synthesis of Adaptive Spline Architectures and Agentic Paradigms in 2026 1. Introduction: The Structural Turn in Deep Learning The trajectory of artificial intelligence research in the mid-2020s has been characterized by a decisive pivot away from the "Depth Hypothesis"âthe long-standing conviction that stacking layers of fixed, node-centric non-linearities (such as Rectified Linear Units or GeLUs) is the singular path to increasing representational power. For nearly a decade, the Multi-Layer Perceptron (MLP) served as the atomic unit of deep learning, embedding a fundamental assumption: that the complexity of the world is best approximated by global linear transformations followed by static point-wise activations. However, the years 2025 and 2026 have witnessed the emergence of a "Structural Turn," a paradigm shift where the focus has moved from the depth of the network to the mathematical quality of the connections themselves. At the forefront of this shift is the Kolmogorov-Arnold Network (KAN), an architecture that relocates learnable non-linearities from the neurons to the edges, parameterizing weights not as scalar values but as univariate B-spline functions. This architectural reorientation is not merely a cosmetic change; it represents a fundamental rethinking of how neural networks approximate continuous functions, grounded in the rigorous mathematical framework of the Kolmogorov-Arnold Representation Theorem of 1957.1 Simultaneously, in the domain of Natural Language Processing (NLP), the limitations of fixed context windows have necessitated a similar structural revolution, giving rise to Recursive Language Models (RLMs) that replace monolithic attention mechanisms with agentic, recursive control flows.3 This report presents an exhaustive technical analysis of these advancements. Unlike standard survey papers, this document prioritizes a "recurse the data" methodology: we do not merely summarize findings but verify the underlying mathematical formulations, cross-reference empirical contradictions, and synthesize second-order insights regarding the causal mechanisms of catastrophic forgetting and context retention. We scrutinize the "Nexus Mirror"âa conceptual framework suggesting that the modular additivity of KANs and the recursive nature of RLMs mirror the causal and physical structures of reality more faithfully than the entangled representations of traditional MLPs.1 By rigorously checking the math of B-spline recursions, least-squares grid extensions, and intrinsic dimensionality bounds, we aim to provide a definitive account of the state of neural architecture in 2026. 2. Theoretical Foundations: The Kolmogorov-Arnold Paradigm To understand the operational mechanics and the theoretical legitimacy of KANs, one must first dissect the mathematical divergence between the original representation theorem proposed in the mid-20th century and its practical realization in modern computational frameworks. 2.1 The Kolmogorov-Arnold Representation Theorem (1957) In 1957, answering David Hilbertâs thirteenth problem, mathematicians Andrey Kolmogorov and Vladimir Arnold established a representation theorem that fundamentally challenged the understanding of multivariate functions. The theorem posits that any continuous multivariate function $f: ^n \to \mathbb{R}$ can be represented as a superposition of continuous univariate functions and addition. The canonical form of this representation is given by: $$f(x_1, \dots, x_n) = \sum_{q=0}^{2n} \Phi_q \left( \sum_{p=1}^{n} \psi_{p,q}(x_p) \right)$$ In this formulation, the inner summation $\sum_{p=1}^{n} \psi_{p,q}(x_p)$ maps the $n$-dimensional input vector to a scalar value, which is then processed by the outer function $\Phi_q$. Crucially, the theorem asserts that the inner functions $\psi_{p,q}$ are continuous and monotonic, and remarkably, they are independent of the target function $f$.2 All information specific to $f$ is encoded in the outer functions $\Phi_q$. Mathematical Verification and Historical Critique: While theoretically profound, the direct application of this theorem to neural networks was stalled for decades by a critical practical limitation. As highlighted by Girosi and Poggio (1989), the inner functions $\psi_{p,q}$ constructed in the original proofs are "pathological"âthey are highly non-smooth, often exhibiting fractal characteristics that make them indistinguishable from noise in a practical setting.8 Because these functions are non-differentiable (or have derivatives that are singular almost everywhere), they are fundamentally incompatible with gradient descent-based learning algorithms like backpropagation. Thus, for nearly seventy years, the Kolmogorov-Arnold theorem was regarded as a mathematical curiosityâan existence proof with no constructive utility for machine learning. 2.2 The Modern KAN Architecture (2024-2026) The breakthrough that enabled the KAN architectures of 2025/2026 did not come from solving the fractal nature of the original $\psi$ functions, but rather from relaxing the theorem's strict conditions. The modern KAN specification, introduced by Liu et al. (2024) and expanded upon in 2025, generalizes the theorem to arbitrary network depths and widths, and most importantly, replaces the fixed, fractal inner functions with learnable, smooth splines.1 A KAN layer in this modern paradigm is defined not by a weight matrix $W$, but by a function matrix $\mathbf{\Phi}$. If a layer has $n_{in}$ inputs and $n_{out}$ outputs, the layer is parameterized by a grid of $n_{in} \times n_{out}$ univariate functions: $$\mathbf{\Phi} = \{ \phi_{q,p} \}, \quad p=1\dots n_{in}, \quad q=1\dots n_{out}$$ The pre-activation of the $q$-th neuron in the subsequent layer is the sum of these function outputs: $$x_{q}^{(l+1)} = \sum_{p=1}^{n_{l}} \phi_{q,p}^{(l)} \left( x_{p}^{(l)} \right)$$ This structure fundamentally differs from the MLP. In an MLP, the linear combination happens before the non-linearity ($ \sigma(\sum w x) $). In a KAN, the non-linearity is applied to each input individually *before* the summation ($\sum \phi(x)$). This "pre-summation non-linearity" allows the network to model complex multiplicative interactions (like $x \times y$) through the identity $xy = \frac{1}{4}[(x+y)^2 - (x-y)^2]$, using only sums and univariate squaresâa capacity that MLPs struggle to achieve without significant depth.1 2.3 Mathematical Verification of B-Splines and Recursion The choice of basis function for $\phi(x)$ is the critical engineering decision in KANs. To enable local plasticityâthe ability to update knowledge in one region of the input space without corrupting knowledge in distant regionsâKANs utilize B-splines. A B-spline curve is constructed from a linear combination of B-spline basis functions $N_{i,k}(x)$ of order $k$: $$\phi(x) = \sum_{i} c_i N_{i,k}(x)$$ The basis functions are defined recursively via the Cox-de Boor formula. We explicitly verify the recursive structure here to confirm the local support property claimed in the literature.13 Base Case ($k=0$): The zeroth-order basis function is a step function (indicator function) over the $i$-th knot interval $$. This mathematical fact is the engine of KANs' continual learning capability: updating a coefficient $c_i$ affects the function $\phi(x)$ only within the compact support of $N_{i,k}(x)$. If a new task provides data outside this interval, the coefficient $c_i$ receives a zero gradient and remains unchanged, thereby preserving the "memory" of the previous task.15 Correction on Notation: Snippets 13 and 14 utilize slightly different indexing conventions ($B_{i,n}$ vs $N_{i,k}$). However, the underlying recurrence relation is identical. It is crucial to note that efficient implementations (like EfficientKAN) assume a uniform grid where $t_{i+1} - t_i = h$ (constant), which simplifies the denominator terms to constants (e.g., $k \cdot h$), replacing division operations with simpler multiplications to accelerate GPU throughput.17 3. Computational Implementation: From PyKAN to MatrixKAN The transition from theoretical construct to practical tool involved significant algorithmic optimization. The initial implementation, referred to as PyKAN, prioritized mathematical clarity over computational efficiency, leading to severe bottlenecks that hindered scaling. 3.1 The Memory Bottleneck in PyKAN In the naive PyKAN implementation 18, the evaluation of spline bases was performed by expanding the input tensor. For a batch size $B$, input dimension $N_{in}$, and grid size $G$, PyKAN would expand the input $x$ to a tensor of shape $(B, N_{in}, G)$. Memory Complexity: $O(B \cdot N_{in} \cdot G)$. Issue: For high-dimensional data (e.g., an image with flattened dimension 1024) and fine grids (e.g., $G=100$), this intermediate tensor becomes prohibitively large, exhausting GPU VRAM even for small batches. 3.2 EfficientKAN: The Matrix Reformulation To address this, the community developed EfficientKAN.17 This implementation reformulates the B-spline computation. instead of expanding the input, it exploits the fact that the spline output is a linear combination of basis functions. Algorithmic Verification: Instead of computing the full expansion, EfficientKAN likely calculates the basis activations $N_{i,k}(x)$ and performs the linear combination with coefficients $c_i$ as a matrix multiplication. Optimization: The memory complexity is reduced to $O(B \cdot N_{in} + N_{in} \cdot N_{out} \cdot G)$ because the batch dimension is decoupled from the grid expansion in memory. Result: Snippet 17 notes that this "simplifies the computation to a basic matrix multiplication." This reformulation was essential for enabling KANs to be used in deeper architectures like Vision Transformers. 3.3 MatrixKAN: Parallelizing the Recursion A further refinement, MatrixKAN, optimizes the Cox-de Boor recursion itself.20 Since t
This study conducts a bibliometric review of Bitcoin research in the Business and Economics domains, using VOSviewer to visualize network structures and Bidirectional Encoder Representations from Transformers Topic (BERTopic) to derive semantically coherent topic clusters. The analysis identifies five major research themes: (1) Diversification, hedging, and safe-haven properties; (2) Market dynamics, efficiency, and investor behavior; (3) Bitcoin price and volatility prediction attempts; (4) Environmental impact of Bitcoin; and (5) Financial impact of Central Bank Digital Currency (CBDC). Based on these themes, the study recommends further investigation into the influence of Exchange-Traded Fund (ETF) approvals, regulatory frameworks, and institutional investor participation on Bitcoinâs safe-haven potential; the role of market dynamics and regulatory interventions; early detection of herding behavior and price bubbles; the integration of machine learning and deep-learning models for price prediction; the environmental costs associated with mining; and the evolving regulatory and implementation challenges of CBDCs. Overall, this review synthesizes existing scholarship and outlines future research directions for the rapidly evolving cryptocurrency ecosystem.
Owen Dugan, Garcia, Roberto, Ronny Junkins, Jerry Liu ¡ 8 authors
The success of large language models (LLMs) can be attributed in part to their ability to efficiently store factual knowledge as key-value mappings within their MLP parameters. Recent work has proposed explicit weight constructions to build such fact-storing MLPs, providing an improved understanding of LLM fact storage mechanisms. In this paper, we introduce an MLP construction framework that improves over previous constructions in three areas: it 1) works for all but a measure-zero set of feasible input-output pairs, 2) achieves asymptotically optimal parameter efficiency matching information-theoretic bounds for some embeddings, and 3) maintains usability within Transformers for factual recall. Through our improvements, we 1) discover a metric on value embeddings that characterizes facts-per-parameter scaling for both constructed and gradient-descent-trained MLPs, 2) identify a simple encoder-decoder mechanism that empirically matches gradient-descent MLP facts-per-parameter asymptotics across all the inputs and outputs we test, and 3) uncover a fundamental tradeoff between an MLP's fact-storage capacity and its usability within Transformers. Finally, we demonstrate a proof-of-concept application of fact-storing MLPs: modular fact editing on one-layer Transformers by \textit{replacing entire MLPs at once}.
Shrutika Singh, Anton Alyakin, Daniel Alexander Alber, Jaden Stryker ¡ 12 authors
The performance of Large Language Models (LLMs) on multiple-choice question (MCQ) benchmarks is frequently cited as proof of their medical capabilities. We hypothesized that LLM performance on medical MCQs may in part be illusory and driven by factors beyond medical content knowledge and reasoning capabilities. To assess this, we created a novel benchmark of free-response questions with paired MCQs (FreeMedQA). Using this benchmark, we evaluated three state-of-the-art LLMs (GPT-4o, GPT-3.5, and LLama-3-70B-instruct) and found an average absolute deterioration of 39.43% in performance on free-response questions relative to multiple-choice (p = 1.3 * 10 -5 ) which was greater than the human performance decline of 22.29%. To isolate the role of the MCQ format on performance, we performed a masking study, iteratively masking out parts of the question stem. At 100% masking, the average LLM multiple-choice performance was 6.70% greater than random chance (p = 0.002) with one LLM (GPT-4o) obtaining an accuracy of 37.34%. Notably, for all LLMs the free-response performance was near zero. Our results highlight the shortcomings in medical MCQ benchmarks for overestimating the capabilities of LLMs in medicine, and, broadly, the potential for improving both human and machine assessments using LLM-evaluated free-response questions.
Open access
Artificial Intelligence in Healthcare and Education
Ke Zhang, Xiaoning Zhao, Chaocheng Zheng, Jiahong Ning ¡ 8 authors
This study proposes Tool-RoCo, a novel benchmark for evaluating large language models (LLMs) in long-term multi-agent cooperation based on RoCo, a multi-robot cooperative benchmark. Recent research on LLM-based multi-agent systems has relied on predefined orchestration, while ignoring agent autonomy. Tool-RoCo treats other agents as tools and introduces cooperative tools, leveraging tool usage to evaluate multi-agent cooperation and self-organization. Tool usage means that each agent (LLM) selects a tool from a candidate set based on the current state, receives feedback, and adjusts its selection in subsequent rounds. To evaluate different autonomy levels, we propose four LLM paradigms: (1) centralized cooperation, where a single LLM allocates tools to all agents; (2) centralized self-organization, where a central LLM autonomously activates agents while keeping others inactive; (3) decentralized cooperation, where each agent has its own LLM and calls tools based on local information; and (4) self-organization, where a randomly chosen initial agent can request collaboration, activating additional agents via tool calls. Tool-RoCo includes three multi-robot tasks, SORT, PACK, and CABINET, to measure format and parameter accuracy and agent coordination through tool usage. The results using several LLMs showed that cooperative tools accounted for only 7.09% of all tools, indicating that LLM-based agents rarely invoked others as assistants. Moreover, activation tools accounted for 96.42%, suggesting that current LLMs tend to maintain active agents while seldom deactivating them for adaptive coordination. Tool-RoCo provides a systematic benchmark to evaluate LLM autonomy and cooperation in multi-agent tasks. Code and Demo: https://github.com/ColaZhang22/Tool-Roco
Web3 technologies, notably Non-Fungible Tokens (NFTs) and Decentralized Finance (DeFi), have generated extensive social media discourse. This study integrates Social Network Analysis (SNA) and BERTopic to examine spreader roles in shaping Web3 conversations on platform X. We collected 12,925 NFT and 7,087 DeFi posts from August 23 to September 23, 2025, using domain-specific keywords. Data preprocessing removed duplicates, spam, and irrelevant content through URL stripping, hashtag filtering, and manual verification of top spreaders to exclude automated accounts. All collection adhered to X's Terms of Service using publicly available English-language posts without personal identifying information. In-degree centrality analysis identified top spreaders. @GiveRep in NFT achieved a time reached of 288 hours with an average propagation speed of two hours, while @BioProtocol in DeFi demonstrated a uniform persistence of 192 hours uniform persistence across top actors. BERTopic analysis revealed thematic differences. NFT discussions centered on community engagement and speculation, whereas DeFi discourse concentrated on protocol infrastructure and yield mechanisms. NFT spreaders exhibited varied influence duration aligned with thematic diversity, while DeFi spreaders showed uniform persistence constrained by content overlap. This integrated framework advances computational social research methodologies and offers practical insights for Web3 stakeholders to identify key influencers and optimize community engagement strategies. Limitations include single-platform focus and one-month observation period.
We step outside the P = NP vs. P â NP dichotomy and, following a co-evolutionary, hypothesis-first program, we frame evidence by the accounting constraint P(L, t) + NP(L, t) = 1, where t indexes registered time windows and L indexes structural layers of analysis. The credit assigned to constructive computation P(L, t) versus certificate-based reasoning NP(L, t) may shift across windows and layers, but their sum is conserved by design. Within this multilayer, time-indexed lens, we propose a test object for proof in 3-SAT: a small, auditable branching set K. Our operational hypothesis is that, within controlled experimental windows, there exists K â V(F) with |K| ⤠c¡log n such that, for every partial assignment Îą: K â {0,1}, the restricted formula F ⣠ι terminates in polynomial time and emits a publicly verifiable certificate (a satisfying assignment or a DRAT/DRUP-style unsatisfiability proof). Because 2^|K| = n^O(1), exhaustive branching over K is polynomial inside the window, enabling artifact-backed constructive behavior without asserting a universal algorithm. We (i) define auditable objects and falsifiable hypotheses, (ii) sketch a Ď-rounds normalization pipeline that contracts structure while logging transformations, (iii) posit a finite catalog of local obstructions with radius-2 witnesses, (iv) outline a greedy hitting-set routine to assemble K, and (v) introduce protection mechanisms against recovery of K by an adversary (commitments and zero-knowledge). Evidence will be supplied via reproducible artifacts (DRAT logs, commitments, run ledgers) and transport tests across registered windows and layers, and will be interpreted under the constraint P(L, t) + NP(L, t) = 1, in a manner consistent with kernelization barriers and sparsification limits.
Fine-tuning large language models (LLMs) is crucial for adapting them to specific tasks, yet it remains computationally demanding and raises concerns about correctness and privacy, particularly in untrusted environments. Although parameter-efficient methods like Low-Rank Adaptation (LoRA) significantly reduce resource requirements, ensuring the security and verifiability of fine-tuning under zero-knowledge constraints remains an unresolved challenge. To address this, we introduce VeriLoRA, the first framework to integrate LoRA fine-tuning with zero-knowledge proofs (ZKPs), achieving provable security and correctness. VeriLoRA employs advanced cryptographic techniques -- such as lookup arguments, sumcheck protocols, and polynomial commitments -- to verify both arithmetic and non-arithmetic operations in Transformer-based architectures. The framework provides end-to-end verifiability for forward propagation, backward propagation, and parameter updates during LoRA fine-tuning, while safeguarding the privacy of model parameters and training data. Leveraging GPU-based implementations, VeriLoRA demonstrates practicality and efficiency through experimental validation on open-source LLMs like LLaMA, scaling up to 13 billion parameters. By combining parameter-efficient fine-tuning with ZKPs, VeriLoRA bridges a critical gap, enabling secure and trustworthy deployment of LLMs in sensitive or untrusted environments.
In the rapidly evolving landscape of digital marketing and electronic commerce, short-form contentâparticularly on platforms like Twitter (now X)âhas become pivotal for real-time branding, community engagement, and product promotion. The rise of Non-Fungible Tokens (NFTs) and Web3 ecosystems further underscores the need for domain-specific, engagement-oriented social media content. However, automating the generation of such content while balancing linguistic quality, semantic relevance, and audience engagement remains a substantial challenge. To address this, we propose RL-TweetGen, a socio-technical framework that integrates instruction-tuned large language models (LLMs) with reinforcement learning (RL) to generate concise, impactful, and engagement-optimized tweets. The framework incorporates a structured pipeline comprising domain-specific data curation, semantic classification, and intent-aware prompt engineering, and leverages Parameter-Efficient Fine-Tuning (PEFT) with LoRA for scalable model adaptation. We fine-tuned and evaluated three LLMsâLLaMA-3.1-8B, Mistral-7B Instruct, and DeepSeek 7B Chatâguided by a hybrid reward function that blends XGBoost-predicted engagement scores with expert-in-the-loop feedback. To enhance lexical diversity and contextual alignment, we implemented advanced decoding strategies, including Tailored Beam Search, Enhanced Top-p Sampling, and Contextual Temperature Scaling. A case study focused on NFT-related tweet generation demonstrated the practical effectiveness of RL-TweetGen. Experimental results showed that Mistral-7B achieved the highest lexical fluency (BLEU: 0.2285), LLaMA-3.1 exhibited superior semantic precision (BERT-F1: 0.8155), while DeepSeek 7B provided balanced performance. Overall, RL-TweetGen presents a scalable and adaptive solution for marketers, content strategists, and Web3 platforms seeking to automate and optimize social media engagement. The framework advances the role of generative AI in digital commerce by aligning content generation with platform dynamics, user preferences, and marketing goals.
Ana-Maria Istrate, Fausto MilletarÏ, Fabrizio Castrotorres, Jakub M. Tomczak ¡ 7 authors
Abstract Reasoning models are typically trained against verification mechanisms in formally specified systems such as code or symbolic math. In open domains like biology, however, we lack exact rules to enable large-scale formal verification and instead often rely on lab experiments to test predictions. Such experiments are slow, costly, and cannot scale with computation. In this work, we show that world models of biology or other prior knowledge can serve as approximate oracles for soft verification , allowing reasoning systems to be trained without additional experimental data. We present two paradigms of training models with approximate verifiers: RLEMF : reinforcement learning with experimental model feedback and RLPK : reinforcement learning from prior knowledge. Using these paradigms, we introduce rbio1 , a reasoning model for biology post-trained from a pretrained LLM with reinforcement learning, using learned biological models for verification during training. We demonstrate that soft verification can distill biological world models into rbio1 , enabling it to achieve state-of-the-art performance on perturbation prediction in the PerturbQA benchmark. We further show that composing multiple AI-verifiers improves performance and that models trained with soft biological rewards transfer zero-shot to cross-domain tasks such as disease-state prediction. We present rbio1 as a proof of concept that predictions from biological models can train powerful reasoning systems using simulations rather than experimental data, offering a new paradigm for model training.
Fatou Ndiaye Mbodji, Mame Marieme C. Sougoufara, WendkÝuni A. M. Christian Ouedraogo, Alioune Diallo ¡ 7 authors
Smart contract comment generation has gained traction as a means to improve code comprehension and maintainability in blockchain systems. However, evaluating the quality of generated comments remains a challenge. Traditional metrics such as BLEU and ROUGE fail to capture domain-specific nuances, while human evaluation is costly and unscalable. In this paper, we present evalSmarT, a modular and extensible framework that leverages large language models (LLMs) as evaluators. The system supports over 400 evaluator configurations by combining approximately 40 LLMs with 10 prompting strategies. We demonstrate its application in benchmarking comment generation tools and selecting the most informative outputs. Our results show that prompt design significantly impacts alignment with human judgment, and that LLM-based evaluation offers a scalable and semantically rich alternative to existing methods.ResourcesVideo Demo: https://youtu.be/HXS_Yiszoz4Code and Data: https://anonymous.4open.science/r/SC_code_summarization-4653
Abderahman Rejeb, Karim Rejeb, Heba F. Zaher, Steve Simske
This paper explores the intersection of blockchain technology and smart cities to support the transition toward decentralized, secure, and sustainable urban systems. Drawing on co-word analysis and BERTopic modeling applied to the literature published between 2016 and 2025, this study maps the thematic and technological evolution of blockchain in urban environments. The co-word analysis reveals blockchainâs foundational role in enabling secure and interoperable infrastructures, particularly through its integration with IoT, edge computing, and smart contracts. These systems underpin critical urban services such as transportation, healthcare, energy trading, and waste management by enhancing data privacy, authentication, and system resilience. The application of BERTopic modeling further uncovers a shift from general technological exploration to more specialized and sector-specific applications. These include real-time mobility systems, decentralized healthcare platforms, peer-to-peer energy exchanges, and blockchain-enabled drone coordination. The results demonstrate that blockchain increasingly supports cross-sectoral innovation, enabling transparency, trust, and circular flows in urban systems. Overall, the current study identifies blockchain as both a technological backbone and an ethical infrastructure for smart cities that supports secure, adaptive, and sustainable urban development.
As large language models (LLMs) are used in sensitive fields, accurately verifying their computational provenance without disclosing their training datasets poses a significant challenge, particularly in regulated sectors such as healthcare, which have strict requirements for dataset use. Traditional approaches either incur substantial computational cost to fully verify the entire training process or leak unauthorized information to the verifier. Therefore, we introduce ZKPROV, a novel cryptographic framework allowing users to verify that the LLM's responses to their prompts are trained on datasets certified by the authorities that own them. Additionally, it ensures that the dataset's content is relevant to the users' queries without revealing sensitive information about the datasets or the model parameters. ZKPROV offers a unique balance between privacy and efficiency by binding training datasets, model parameters, and responses, while also attaching zero-knowledge proofs to the responses generated by the LLM to validate these claims. Our experimental results demonstrate sublinear scaling for generating and verifying these proofs, with end-to-end overhead under 3.3 seconds for models up to 8B parameters, presenting a practical solution for real-world applications. We also provide formal security guarantees, proving that our approach preserves dataset confidentiality while ensuring trustworthy dataset provenance.
Purpose Cryptocurrencyâs novelty and volatilityâcombined with the absence of standardized reporting prior to 2023âcreated an opaque information environment. This study explores whether such conditions enabled assertive impression management in corporate reporting. We examine how firms not only varied the volume of cryptocurrency disclosures over time, but also strategically manipulated their readability . Additionally, we use this context to demonstrate the utility of machine learning and natural language processing tools for consistent analysis of complex financial narratives. Study design We analyze full-text annual reports, MD&A sections, and proxy statements from five publicly traded U.S. firms with diverse cryptocurrency involvements. Our methodology includes machine learning-based topic modeling, readability assessment using standardized indices, and visualization tools. Findings (i) Information Demand: Google search trends for target firms are strongly associated with Bitcoin price movements, reflecting external attention cycles. (ii) Impression Management: Firms increase both the frequency and readability of crypto disclosures in favorable markets and reduce or obscure them in downturns, consistent with strategic impression management. (iii) Readability: Crypto-related disclosures are significantly more readable than non-crypto sections from the same reports suggesting deliberate simplification. Contributions This study advances the limited literature on cryptocurrency disclosure by offering a textual and behavioral lens on corporate impression management. A key contribution is the integration of readability metrics, public attention signals, and NLP tools into disclosure analysis. We highlight how firms use both narrative framing and readability engineering as tools to influence perceptionâespecially in periods of regulatory uncertainty. Implications Our findings have direct implications for policy and practice: (i) Policymakers should consider not only disclosure quantity but also its linguistic clarity and comparability, especially for volatile assets. (ii) Investors and analysts can use automated text analysis to detect subtle impression management tactics and to interpret the strategic use of clarity in disclosure narratives.
Philip Kang, Yong Jae Ko, Sangpil Youm, Yoonki Chun ¡ 6 authors
Purpose The current study attempted to unravel the complexities of sport Non-Fungible Token (NFT) consumption by exploring consumers' perceived values and risks of sport NFTs. Design/methodology/approach To achieve this goal, the authors conducted an exploratory study by collecting 23,445 tweets from Twitter and adopted artificial intelligence tools (i.e. latent Dirichlet allocation topic modeling and visualization, Word2Vec and uniform manifold approximation and projection) to analyze the data. Findings The results revealed that sport NFT platforms (i.e. NBA Top Shot, NFL All Day and UFC Strike) engage users through a blend of intrinsic and extrinsic motivations, including personal enjoyment, financial rewards and social recognition. Originality/value The current research not only illuminates the newly emerged NFT market but also provides exploratory evidence to understand sport NFT consumersâ value perception by using machine and deep learning approaches. Practically, the current study provides valuable insights to sport NFT brand managers by identifying significant values that trigger sport NFT consumers' engagement with sport NFT consumption behavior.
Md. Nahidul Islam Opu, Md Shahidul Islam, Sara Rouhani, Shaiful Chowdhury
Blockchain-based software systems are increasingly deployed across diverse domains, yet a systematic understanding of their development challenges remains limited. This paper presents a large-scale empirical study of 497,742 issues mined from 1,209 open-source blockchain projects hosted on GitHub. Employing BERTopic, a transformer-based topic modeling technique, we identify 49 distinct issue topics and organize them hierarchically into 11 major subcategories. Our analysis reveals that both general software development issues and blockchain-specific concerns are nearly equally represented, with Wallet Management and UI Enhancement emerging as the most prominent topics. We further examine the temporal evolution of issue categories and resolution times, finding that Wallet issues not only dominate in frequency but also exhibit the longest resolution time. Conversely, Mechanisms issues are resolved significantly faster. Issue frequency surged after 2016 with the rise of Ethereum and decentralized applications, but started declining after 2022. These findings enhance our understanding of blockchain software maintenance, informing the development of specialized tools and practices to improve robustness and maintainability.
Che, Zheng, Taoyu Li, Meng Shen, Hanbiao Du ¡ 5 authors
The untraceability of transactions facilitated by Ethereum mixing services like Tornado Cash poses significant challenges to blockchain security and financial regulation. Existing methods for correlating mixing accounts suffer from limited labeled data and vulnerability to noisy annotations, which restrict their practical applicability. In this paper, we propose StealthLink, a novel framework that addresses these limitations through cross-task domain-invariant feature learning. Our key innovation lies in transferring knowledge from the well-studied domain of blockchain anomaly detection to the data-scarce task of mixing transaction tracing. Specifically, we design a MixFusion module that constructs and encodes mixing subgraphs to capture local transactional patterns, while introducing a knowledge transfer mechanism that aligns discriminative features across domains through adversarial discrepancy minimization. This dual approach enables robust feature learning under label scarcity and distribution shifts. Extensive experiments on real-world mixing transaction datasets demonstrate that StealthLink achieves state-of-the-art performance, with 96.98\% F1-score in 10-shot learning scenarios. Notably, our framework shows superior generalization capability in imbalanced data conditions than conventional supervised methods. This work establishes the first systematic approach for cross-domain knowledge transfer in blockchain forensics, providing a practical solution for combating privacy-enhanced financial crimes in decentralized ecosystems.
This study critically reviews the literature on metaverse technologies, developing an integrative framework to explore their sector-specific implications and transformative impact on business management. Employing the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) framework and machine learning-based BERTopic modeling, the study identifies nine key themes, reflecting the diverse ways augmented reality (AR), virtual reality (VR), extended reality (XR), digital twins, and decentralized finance (DeFi) influence industries. These themes include the metaverse as a tool for economic and environmental policy experiments, navigating financial risk and regulatory dynamics, adapting human resource development to VR-driven environments, Industry 4.0 applications of VR and digital twins, digital twin applications in manufacturing and supply chain optimization, AR and VR in digital marketing and customer experience, AR in enhancing retail and consumer experiences, exploring user interaction and affordances in the metaverse, and VR and AR in tourism experience and engagement. The framework highlights drivers, constraints, and cross-sector linkages, addressing practical challenges such as high implementation costs, regulatory uncertainties, interoperability barriers, cybersecurity risks, and ethical concerns surrounding data privacy and inclusion. The study critically evaluates contradictions in metaverse adoption, such as the tension between sustainability goals and energy-intensive technologies like blockchain, the gap between immersive training potential and workforce adaptation challenges, and the disparity between metaverse-driven economic models and real-world policy implementation hurdles. Research propositions suggest integrating metaverse technologies into business operations while balancing ethical dimensions, psychological impacts, cost limitations, and accessibility barriers. Additionally, the study advocates for expanding theoretical frameworks such as the Resource-Based View (RBV), Technology Acceptance Model (TAM), and experiential learning to account for the dynamic capabilities, risks, and industry-specific constraints of metaverse adoption. Policymakers and practitioners are encouraged to address regulatory and ethical challenges, sectoral disparities, and the unintended consequences of metaverse-driven digital transformation, ensuring operational efficiency, resilience, and consumer engagement while fostering sustainable and inclusive adoption. This research offers actionable insights for strategic implementation, interdisciplinary theoretical expansion, and ethical progress in business management.
The rapid advancement of Large Language Models (LLMs) has catalyzed the development of multi-agent systems, where multiple LLM-based agents collaborate to solve complex tasks. However, existing systems predominantly rely on centralized coordination, which introduces scalability bottlenecks, limits adaptability, and creates single points of failure. Additionally, concerns over privacy and proprietary knowledge sharing hinder cross-organizational collaboration, leading to siloed expertise. To address these challenges, we propose AgentNet, a decentralized, Retrieval-Augmented Generation (RAG)-based framework that enables LLM-based agents to autonomously evolve their capabilities and collaborate efficiently in a Directed Acyclic Graph (DAG)-structured network. Unlike traditional multi-agent systems that depend on static role assignments or centralized control, AgentNet allows agents to specialize dynamically, adjust their connectivity, and route tasks without relying on predefined workflows. AgentNetâs core design is built upon several key innovations: (1) Fully Decentralized Paradigm: Removing the central orchestrator, allowing agents to coordinate and specialize autonomously, fostering fault tolerance and emergent collective intelligence. (2) Dynamically Evolving Graph Topology: Real-time adaptation of agent connections based on task demands, ensuring scalability and resilience. (3) Adaptive Learning for Expertise Refinement: A retrieval-based memory system that enables agents to continuously update and refine their specialized skills. By eliminating centralized control, AgentNet enhances fault tolerance, promotes scalable specialization, and enables privacy-preserving collaboration across organizations. Through decentralized coordination and minimal data exchange, agents can leverage diverse knowledge sources while safeguarding sensitive information. Experimental results demonstrate that AgentNet outperforms traditional centralized multi-agent systems, significantly improving efficiency, adaptability, and scalability in dynamic environments, making it a promising foundation for next-generation autonomous, privacy-respecting multi-agent ecosystems.
In many-task optimization scenarios, surrogate models are valuable for mitigating the computational burden of repeated fitness evaluations across tasks. This study proposes a novel meta-surrogate framework to assist many-task optimization, by leveraging the knowledge transfer strengths and emergent capabilities of large language models (LLMs). We formulate a unified framework for many-task fitness prediction, by defining a universal model with metadata to fit a group of problems. Fitness prediction is performed on metadata and decision variables, enabling efficient knowledge sharing across tasks and adaptability to new tasks. The LLM-based meta-surrogate treats fitness prediction as conditional probability estimation, employing a unified token sequence representation for task metadata, inputs, and outputs. This approach facilitates efficient inter-task knowledge sharing through shared token embeddings and captures complex task dependencies via multi-task model training. Experimental results demonstrate the model's emergent generalization ability, including zero-shot performance on problems with unseen dimensions. When integrated into evolutionary transfer optimization (ETO), our framework supports dual-level knowledge transfer -- at both the surrogate and individual levels -- enhancing optimization efficiency and robustness. This work establishes a novel foundation for applying LLMs in surrogate modeling, offering a versatile solution for many-task optimization.
The Shakti series of 100M, 250M, and 500M models offers compact, resource-efficient language models designed for edge AI deployment. Unlike large models like GPT-3 and LLaMA that demand cloud-based infrastructure, Shakti models operate seamlessly on low-resource devices, including smartphones, smart TVs, IoT systems, drones, and low-end GPUs. They ensure minimal energy consumption, privacy-preserving computation, and real-time performance without internet dependency. Optimized for efficiency, Shakti models come in quantized versions (int8, int5, int4) for even faster, lighter execution on edge devices. The 2.5B Shakti model has demonstrated strong performance while maintaining low latency, paving the way for the smaller, highly efficient 100M, 250M, and 500M models. Built on Responsible AI principles, Shakti prioritizes fairness, transparency, and trust while mitigating risks such as bias, privacy concerns, and high carbon footprints. These models are ideal for sensitive domains like finance, healthcare, and legal services, providing cost-effective, sustainable, and scalable AI solutions with on-device data security. Each model is tailored for specific applications. Shakti-100M excels in text generation, summarization, and chatbots for IoT and mobile apps. Shakti-250M specializes in domain-specific tasks such as contract analysis and personalized financial or healthcare advice. Shakti-500M, a versatile model, enhances customer support, content creation, and virtual assistants with multilingual capabilities and long-context understanding. By decentralizing AI, the Shakti series democratizes access to intelligent, ethical, and impactful AI solutions across industries.
Blockchain technologies are increasingly recognized as a transformative force across industries, offering potential solutions for information management, data security, and operational efficiency improvement. However, integration into the healthcare sector faces significant barriers, ranging from technical challenges to organizational resistance. In this study, a methodology is proposed that examines these critical barriers through a comprehensive analysis of 3265 academic papers and derives actionable solutions from 1566 patents published between 2016 and 2023. This approach bridges the gap between identifying challenges and implementing solutions. Using the âStepwise Weight Assessment Ratio Analysis (SWARA)â, twelve critical adoption challenges are analyzed, while Top2Vec-based topic modeling identifies innovations that best address the ranked barriers. In addition to this, the proposed âhealthcare ecosphereâ knowledge map serves as a comprehensive tool to analyze key stakeholders, their interactions, and the alignment of adoption barriers in solution spaces. The findings show that innovations in blockchain technologies are heavily concentrated in areas such as data security and application functionalities, whereas other critical domainsâsuch as consensus mechanisms, governance, and regulatory frameworksâremain underexplored, pointing to opportunities for growth and development. The mapping of barriers to solutions provides practical guidance for healthcare providers, policymakers, and technologists seeking to implement the blockchain technologies effectively. ⢠Identified and ranked twelve key adoption barriers from technological and social contexts using SWARA methodology. ⢠Analyzed trends in blockchain patent grants to strategically address the barriers. ⢠Patent documents are clustered and mapped to address barriers using machine learning approach through Top2Vec. ⢠Noting patent grants are concentrated in specific areas of applications. ⢠Highlighted opportunities for innovation in under-explored blockchain technology layers.
The growing power of large language models (LLMs) has revolutionized how people access and utilize information. Notably, the LLMs excel at performing fine-grained data representation, which facilitates precise retrieval of information. They also generate high-quality answers based on external references, enabling the production of useful knowledge. The recent introduction of reasoning models, like OpenAI O1 and DeepSeek R1, marks another leap forward, highlighting LLMs' ability to think progressively before delivering final answers. This breakthrough significantly improves the ability to address complex tasks, e.g., coding and math proofs. Inspired by this progress, we aim to develop similar capabilities for retrieval models, which hold great promise for tackling critical challenges in the field, including multi-task retrieval, zero-shot retrieval, and tasks requiring intensive reasoning of complex relationships. With this motivation, we propose a novel approach called O1 Embedder, which generates useful thoughts for the input query before making retrieval for the target documents. To realize this objective, we conquer two technical difficulties. First, we design a data synthesis workflow, creating training signals for O1 Embedder by generating initial thoughts from an LLM-expert and subsequently refining them using a retrieval committee. Second, we optimize the training process, enabling a pre-trained model to be jointly fine-tuned to generate retrieval thoughts via behavior cloning and perform dense retrieval through contrastive learning. Our approach is evaluated by comprehensive experiments, where substantial improvements are achieved across 12 popular datasets, spanning both in-domain and out-of-domain scenarios. These results highlight O1 Embedder's remarkable accuracy and generalizability, paving the way for the development of next-generation IR foundation models.