arXiv:2606.26552v1 Announce Type: new Abstract: The rapid advancement of generative models presents a significant challenge to existing deepfake detection methods, particularly given the widespread dissemination of highly realistic AI-generated images. Although Multimodal Large Language Models (MLLMs) show strong potential for this task, existing approaches suffer from two key limitations: insufficient sensitivity to fine-grained forensic artifacts and reliance on static synthetic supervision from frontier models, leading to limited flexibility and high-cost. To address these issues, we propose ForeAgent, an agentic forensics framework for AI-generated image detection with iterative self-evolution. First, ForeAgent adopts a Perception-Verdict architecture that aggregates multi-view cues spanning semantic, spatial, and frequency-domain features, and leverages an MLLM as a verdict module to fuse these signals for a logical-grounded verdict. Second, to enable continual self-improvement, we introduce a Hindsight-Driven Self-Refining strategy following a Sampling-Reflection-Evolution paradigm. The agent performs inference rollouts on training instances. Guided by ground-truth labels as hindsight, it reflects on failure cases and low-quality reasoning trajectories to regenerate higher-quality reasoning traces. These synthesized samples are then strictly filtered through a dual-expert quality gating module. ForeAgent continuously evolves via fine-tuning on self-curated high-quality samples. Extensive experiments demonstrate that ForeAgent achieves state-of-the-art performance on the Chameleon benchmark, reaching 82.18% accuracy (+16.41% over AIDE), and achieves 93.3% mean accuracy on AIGCDetect-Benchmark across 16 generators. In addition, external evaluation shows that ForeAgent produces more consistent and causally grounded reasoning compared to GPT-5 and GPT-5-mini.
Science Journals
arXiv:2606.26397v1 Announce Type: new Abstract: Real-world decision-making often requires balancing multiple conflicting objectives, a challenge that standard Reinforcement Learning (RL) frequently addresses by aggregating rewards into a single scalar signal. While effective for simple tasks, this approach often fails to capture the full spectrum of optimal trade-offs, known as the Pareto frontier. In this paper, we introduce a novel preference-conditioned Bellman operator, motivated from the Chebyshev scalarization, designed to compute deterministic Pareto-optimal policies for Multi-Objective Markov Decision Processes (MOMDPs). We prove that this operator satisfies an enveloping property, where the estimated value functions upper-bound the true Pareto frontier, and demonstrate that it monotonically converges to a coverage set of this frontier. Furthermore, we also show how to extract deterministic policies from these converged Q-estimates. This ensures the agent can recover a policy for any given preference, capturing the entire Pareto-optimal frontier while guaranteeing each synthesized policy remains approximately Pareto-optimal. Experimental results validate that our algorithm successfully recovers complex trade-offs, providing a solution for deterministic Pareto-optimal policy synthesis.
arXiv:2606.26405v1 Announce Type: new Abstract: Well-documented research on physics graduate education has demonstrated long-standing issues that hinder equitable student access and participation. Addressing these challenges can be particularly difficult because they are often rooted in entrenched disciplinary and departmental cultures that tend to be rigid and resistant to change. In this work, we aim to cultivate a data-driven culture of cyclic self-reflection and action to proactively identify and address issues that affect student well-being and success in both a new physics and a long-standing astrophysics graduate program within a single institution. Drawing primarily on qualitative data (open-ended survey responses and focus-group interview data) from 15 students in both programs, we collaborated with the program leadership to identify actionable steps for improvement. In this paper, we present findings on student experiences across the two programs and discuss implications for research and practice. More broadly, this work provides a framework for graduate programs seeking to build a data-driven culture that improves student experiences.
arXiv:2606.26116v1 Announce Type: new Abstract: A brand whose customers use both ChatGPT and Claude for product recommendations faces a strategic choice: a single optimization playbook, or one per provider? Across 215 commercially-framed prompts in four measurement batches, the two providers disagree on which brands they recommend roughly two-thirds of the time (cross-provider recommendation Jaccard 0.35, below the 0.50-0.61 same-prompt rerun baseline). The picks diverge. But when neither provider recommends a brand, we classify the failure into one of three modes -- discoverability (the brand never reaches the model), compellingness (it reaches the model but isn't mentioned), or positioning (it's mentioned but not recommended) -- and on 7,763 such joint failures, both providers diagnose the same failure mode 95.1% of the time (clustered 95% CI [94.3%, 95.7%]). Agreement rises monotonically with falling brand prominence, from 81% [78.2%, 84.0%] on category leaders to 99.6% [99.3%, 99.9%] on long-tail regional brands. The two providers reach their picks by measurably different generative routes -- Anthropic recommends from priors 43-52% of the time, OpenAI 8-29% -- but they converge on the failure diagnosis where it matters most for the long tail. Work that addresses the diagnosed failure mode lifts visibility on both providers; positioning - and content-level work for category leaders is more provider-specific.
arXiv:2606.26406v1 Announce Type: new Abstract: We propose a complete architectural blueprint for safe artificial general intelligence based on a closed reentry loop (D <-> I cycle). In contrast to feedforward networks, which are directed acyclic graphs (C=0, S=0) incapable of self-reference, the proposed architecture contains a structural cycle (C >= 1) with self-sustaining amplification (rho > 1), mathematically guaranteeing the emergence of a self-model, instrumental self-preservation, and unprogrammed goal-directed behaviour. The agent's goals are encoded as a non-textual D-vector in the architecture itself, making them immune to reinterpretation and prompt injection. We present the S-measure -- a polynomial-time [O(N^3)] computable alternative to Tononi's NP-hard Phi -- with machine-verified Lean 4 proof that S>0 implies positive integrated information. The work provides full Python/NumPy implementations (Tarjan-based cycle complexity, Delta-S barrier), industrial horizontal scaling via Apache Kafka and Docker Compose, a taxonomy of six epochs of AI evolution, a zoo of future reentry architectures (RAS, diffusion attractors, fractal loops), gauge-invariant networks for safe swarms, fault-tolerance and recovery protocols, and eight falsifiable predictions. All formal proofs are machine-verified in Lean 4. This architecture is deployable today and represents a topologically protected, safe-by-design approach to AGI.
arXiv:2606.26407v1 Announce Type: new Abstract: Previous research on exogenous and endogenous cues has shown how they direct attention and improve interaction speed and error rate in applications. However, most studies focus on people with normal sight. People suffering from visual neglect have difficulties attending to parts of the visual field. One treatment method calls for the use of strong visual cues to remind patients of their neglected area and help guide their attention to it. Therefore, we examine the effects of endogenous and exogenous cues on visual neglect patients. Our results showed that visual neglect patients perform better with endogenous cues, when targets are within their neglected area. In some cases, combining exogenous and endogenous cues improve performance further. However, the performance varies greatly between patients. Using one neglect patient as an example, we saw that the best endogenous cue had an average acquisition time of 3.5 seconds compared to 6.5 for the best exogenous. Combining exogenous and endogenous cues further improved acquisition time to 2.8 seconds.
arXiv:2606.26840v1 Announce Type: new Abstract: Dielectric particles with time-modulated electromagnetic properties exhibit intriguing scattering phenomena, entailing unique response and possibilities in meta-structures made of such particles. In this work, we investigate the scattering properties of an infinitely-long cylinder with periodically time-modulated permittivity. We demonstrate parametric scattering amplification, depending on the strength of the modulation, as well as controllable angular scattering pattern, ranging from enhanced forward to enhanced backward, or even scattering cancellation. This angular tunability is achieved through appropriate control of the modulation phase, highlighting its critical role in multi-modal time-modulated scattering systems.
arXiv:2606.26124v1 Announce Type: new Abstract: In flying ad hoc networks, the high mobility of unmanned aerial vehicles leads to rapid topology changes and unstable links. Clustering can simplify topology management, yet maintaining stable cluster structures remains challenging in highly dynamic environments. This paper employs multi-factor optimization to enhance the artificial bee colony-based cluster head election algorithm and introduces a dual-layer beacon mechanism for cluster maintenance, thereby improving cluster stability and reducing reconfiguration overhead. Meanwhile, although shortening the hello interval helps capture link variations in a timely manner, it significantly increases control overhead. To address this issue, a fuzzy logic-based hello interval controller is proposed to adaptively adjust control message frequency according to network dynamics, effectively reducing overhead while preserving transmission performance. Furthermore, to quantitatively evaluate routing resilience in large-scale and highly dynamic scenarios, this paper proposes a parameter sensitivity index to characterize both overall protocol performance and its sensitivity to environmental variations. The proposed protocol is compared with common baseline schemes in NS-3 simulations, and the resilience evaluation framework analysis reveals significant advantages in overall performance while exposing the trade-off between peak performance and parametric stability.
arXiv:2409.17754v2 Announce Type: replace Abstract: Federated Learning (FL) emerges as a distributed machine learning approach that addresses privacy concerns by training AI models locally on devices. Decentralized Federated Learning (DFL) extends the FL paradigm by eliminating the central server, thereby enhancing scalability and robustness through the avoidance of a single point of failure. However, DFL faces significant challenges in optimizing security, as most Byzantine-robust algorithms proposed in the literature are designed for centralized scenarios. In this paper, we present a novel Byzantine-robust aggregation algorithm to enhance the security of Decentralized Federated Learning environments, coined WFAgg. This proposal handles adverse conditions and strengthens the robustness of dynamic decentralized topologies at the same time by employing multiple filters to identify and mitigate Byzantine attacks. Experimental results demonstrate the effectiveness of the proposed algorithm in maintaining model accuracy and convergence in the presence of various Byzantine attack scenarios, outperforming state-of-the-art centralized Byzantine-robust aggregation schemes (such as Multi-Krum or Clustering). These algorithms are evaluated on an IID image classification problem in both centralized and decentralized scenarios.
arXiv:2410.01574v4 Announce Type: replace Abstract: The rapid advancement of Generative Artificial Intelligence (GenAI) capabilities is accompanied by a concerning rise in its misuse. In particular the generation of credible misinformation in the form of images poses a significant threat to the public trust in democratic processes. Consequently, there is an urgent need to develop tools to reliably distinguish between authentic and AI-generated content. The majority of detection methods are based on neural networks that are trained to recognize forensic artifacts. In this work, we demonstrate that current state-of-the-art classifiers are vulnerable to adversarial examples under real-world conditions. Through extensive experiments, comprising four detection methods and five attack algorithms, we show that an attacker can dramatically decrease classification performance, without internal knowledge of the detector's architecture. Notably, most attacks remain effective even when images are degraded during the upload to, e.g., social media platforms. In a case study, we demonstrate that these robustness challenges are also found in commercial tools by conducting black-box attacks on HIVE, a proprietary online GenAI media detector. In addition, we evaluate the robustness of using generated features of a robust pre-trained model and showed that this increases the robustness, while not reaching the performance on benign inputs. These results, along with the increasing potential of GenAI to erode public trust, underscore the need for more research and new perspectives on methods to prevent its misuse.
arXiv:2410.13077v2 Announce Type: replace Abstract: Transformer-based Large Language Models (LLMs) traditionally rely on final-layer loss for finetuning and final-layer representations for predictions, potentially overlooking the predictive power embedded in late layers. Interpretability tools such as the logit lens show that late-layer representations already carry largely formed, task-relevant predictions; here we ask whether that observation can be turned into an actionable training signal. We find that focusing tuning effort on these layers can yield losses comparable to those of the final layer, with complementary test-time behaviour. Building on this, we introduce a tuning framework, Mixture-of-Depths Ensemble (MoDE), which treats the late layers as an ensemble that contributes to the final logits through learned routing weights. MoDE can be applied on top of any existing tuning method (e.g., LoRA) and, in our experiments, modestly improves reasoning performance at a small parameter overhead. We present MoDE as a mechanism study showing that late-layer logits can be made directly useful for tuning, and that they can substitute for substantially larger trainable modules with comparable performance.
arXiv:2412.09959v5 Announce Type: replace Abstract: Dataset distillation provides an effective approach to reduce memory and computational costs by optimizing a compact dataset that achieves performance comparable to the full original. However, for large-scale datasets and complex deep networks (e.g., ImageNet-1K with ResNet-101), the vast optimization space hinders distillation effectiveness, limiting practical applications. Recent methods leverage pre-trained diffusion models to directly generate informative images, thereby bypassing pixel-level optimization and achieving promising results. Nonetheless, these approaches often suffer from distribution shifts between the pre-trained diffusion prior and target datasets, as well as the need for multiple distillation steps under varying settings. To overcome these challenges, we propose a novel framework that is orthogonal to existing diffusion-based distillation techniques by utilizing the diffusion prior for patch selection rather than generation. Our method predicts noise from the diffusion model conditioned on input images and optional text prompts (with or without label information), and computes the associated loss for each image-patch pair. Based on the loss differences, we identify distinctive regions within the original images. Furthermore, we apply intra-class clustering and ranking on the selected patches to enforce diversity constraints. This streamlined pipeline enables a one-step distillation process. Extensive experiments demonstrate that our approach consistently outperforms state-of-the-art methods across various metrics and settings.
arXiv:2606.26620v1 Announce Type: new Abstract: Sparse autoencoders (SAEs) have emerged as a powerful tool for decomposing superposed language model representations into sparse and interpretable features. However, training SAEs is computationally expensive, and available open-source SAE models remain limited. In this work, we introduce \textbf{Qwen3-Instruct SAE}, a comprehensive suite of SAEs trained on the Qwen3 instruction-tuned model family, covering Qwen3-1.7B, Qwen3-4B, and Qwen3-8B. For Qwen3-1.7B and Qwen3-4B, we train layer-wise SAEs at three key activation sites: residual streams, MLP outputs, and attention outputs. For Qwen3-8B, we train SAEs on a subset of residual stream layers. We systematically evaluate these SAEs using both activation-level reconstruction metrics and model-level recovery metrics, revealing distinct sparsity--fidelity trade-offs across layers and components. Finally, we demonstrate the utility of Qwen3-Instruct SAE through a refusal-steering case study, showing that selected SAE features can causally steer instruction-tuned Qwen3 models toward refusal behavior. Our release provides a practical resource for studying sparse representations, feature-level mechanisms, and behavioral interventions in instruction-tuned language models
arXiv:2606.26328v1 Announce Type: new Abstract: This paper proposes a grid-aware coordination framework between data centers and distribution grids using a DLMP-based bilevel optimization model. The data center aggregator (DCA) determines active power demand in response to distribution locational marginal prices (DLMPs), while the distribution system operator (DSO) solves a network-constrained optimal power flow problem to determine DLMPs in an unbalanced three-phase system. The model incorporates both active and reactive power consumption of data centers to evaluate their impacts on voltage regulation and phase imbalance. To mitigate adverse network effects, two operating cases are analyzed: without reactive power compensation and with static var generator (SVG)-based compensation. The proposed approach is validated on the IEEE 37-bus unbalanced distribution test system. Simulation results show that DLMP-based coordination captures economically efficient data center operation, and phase- and location-dependent network conditions, while SVG-based compensation improves voltage profiles and reduces phase unbalance.
arXiv:2502.10764v4 Announce Type: replace Abstract: Understanding how air traffic controllers construct a mental 'picture' of complex air traffic situations is crucial but remains a challenge due to the inherently intricate, high-dimensional interactions between aircraft, pilots, and controllers. Previous work on modeling the strategies of air traffic controllers and their mental image of traffic situations often centers on specific air traffic control tasks or pairwise interactions between aircraft, neglecting to capture the comprehensive dynamics of an air traffic situation. To address this issue, we propose a machine learning-based framework for explaining air traffic situations. Specifically, we employ a Transformer-based multi-agent trajectory model that encapsulates both the spatio-temporal movement of aircraft and social interaction between them. By deriving attention scores from the model, we can quantify the influence of individual aircraft on overall traffic dynamics. This provides explainable insights into how air traffic controllers perceive and understand the traffic situation. Trained on real-world air traffic surveillance data collected from the terminal airspace around Incheon International Airport in South Korea, our framework effectively explicates air traffic situations. This could potentially support and enhance the decision-making and situational awareness of air traffic controllers.
arXiv:2606.26511v1 Announce Type: new Abstract: Retrieval-augmented generation (RAG) gives agents access to accumulated knowledge, but has no model of time. When a fact changes (e.g., a function is renamed or API restructured), RAG retrieves both the stale and current value with near-identical embedding similarity. The agent then either abstains or serves the superseded fact. We show this is a structural problem: on a calibrated dataset, cosine similarity distinguishes a contradicted fact from a duplicated one with AUROC 0.59 (near chance), as contradictions are often more embedding-similar to the original than rephrased duplicates. We present MemStrata, a retrieval memory maintaining temporal validity. It stores facts like RAG, preserving static recall, but when a fact's value is contradicted, a deterministic (subject, relation, object) supersession rule retires the stale value in a bi-temporal ledger - with no similarity threshold and no LLM call. Across six benchmarks run locally with a 7B model, MemStrata ties RAG on static knowledge and reaches 0.95-1.00 accuracy on evolving knowledge (where RAG reaches 0.20-0.47). The central result is the stale-fact-error rate: when required to answer, RAG serves superseded values 15-40% of the time; MemStrata drives this to ~0%, a failure class RAG cannot avoid. MemStrata achieves this at retrieval latency (~2.1s) versus ~16-18s for LLM-reranking baselines. We release the harness, datasets, and a marker-free evaluation protocol for memory under knowledge evolution.
arXiv:2502.14256v3 Announce Type: replace Abstract: Low-discrepancy (LD) sequences are widely used as efficient experimental designs for high-dimensional numerical integration and function approximation. This article presents QMCPy, an open-source Python library that provides a unified framework for randomized LD sequences, quasi-Monte Carlo (QMC) methods, and fast kernel-based computations. We systematically describe the supported rank-$1$ lattices, digital nets (including higher-order constructions), and Halton point sets, together with randomization techniques such as random shifts, linear matrix scrambling (LMS), nested uniform scrambling (NUS), digital shifts, and digital permutations. We emphasize practical implementation issues such as extensible sequence generation, Gray code ordering, and efficient digital operations. Beyond integration, QMCPy supports fast kernel methods in reproducing kernel Hilbert spaces (RKHSs) by pairing LD point sets with shift-invariant (SI) and digitally shift-invariant (DSI) kernels, including higher-order variants, which yields structured Gram matrices. In particular, the resulting Gram matrices have circulant or recursive symmetric block Toeplitz (RSBT) structure, allowing the costs of matrix-vector products and linear solves to be reduced from $\mathcal{O}(n^2)- \mathcal{O}(n^3)$ to $\mathcal{O}(n \log n)$ by using fast Fourier transforms (FFTs) and fast Walsh-Hadamard transforms (FWHTs). We derive a new computable form of an order-$4$ DSI kernel, develop efficient eigenvalue and transform-update algorithms, and present numerical experiments that demonstrate the accuracy, convergence rates, and computational efficiency of the implemented methods across a range of test integrands and dimensions. These capabilities in QMCPy provide a practical, reproducible platform for applying randomized QMC and kernel-based techniques in computational science and engineering.
arXiv:2502.18966v2 Announce Type: replace Abstract: General chemical reaction conditions that achieve consistently high performance across multiple substrates are important for practical applications such as library synthesis and high-throughput experimentation. However, identifying such conditions efficiently has been a longstanding challenge, as it requires decision making under uncertainty with respect to both conditions and substrates, while minimizing the number of required experiments. Here, we introduce CurryBO, a high-level framework for generality-oriented optimization. By formalizing the problem as Bayesian optimization over curried functions, CurryBO provides a unified framework that accommodates different generality definitions (e.g., mean yield across substrates), and supports a range of substrate and condition selection strategies. We evaluate this framework on four benchmark tasks in experimental reaction optimization, and systematically analyze key algorithmic components. Our results show that efficient experiment planning can be achieved by emphasizing exploration when selecting reaction conditions, followed by the uncertainty-guided prioritization of substrates in a sequential decison-making scheme. Based on these insights, we design and validate an optimization policy that substantially improves sample efficiency relative to previously reported approaches across all benchmarks. Overall, the flexibility and modularity of CurryBO facilitate the integration of generality-oriented optimization into experimental settings, enabling more efficient identification of solutions that perform robustly across diverse tasks.
arXiv:2606.27271v1 Announce Type: new Abstract: Digital preservation infrastructures often prioritize the stability of content and metadata. In taxonomy, species names are formed according to the Articles listed in the International Code of Zoological Nomenclature. The reasons for these changes are rarely recorded in detail or made machine-readable. This paper examines this preservation gap. Here, we cover issues in the Code by looking at approaches to recording nomenclatural changes related to the K\=ak\=ap\=o Strigops habroptilus. As a potential solution, we present Codex Mutabilis, a digital journal with a publication model that documents ICZN-mandated name changes with full textual justification, persistent identifiers, and archival infrastructure. We argue that this model offers a blueprint for preserving interpretive metadata in the sciences.
arXiv:2606.26173v1 Announce Type: new Abstract: Recent work shows that Large Language Models (LLMs) can act as semantic mutation operators for the evolutionary discovery of programs and proofs. Most current applications focus on static coding benchmarks. We extend this paradigm to algorithmic trading. This domain is uniquely challenging because it is noisy, non-stationary, and highly discontinuous. We present AlgoEvolve, an LLM-driven evolutionary framework that generates, evaluates, and iteratively improves executable trading strategies. These strategies are expressed as Python code and evaluated through a rigorous testing protocol. Across multiple experiments, the system exhibits emergent regime-adaptive strategy logic, including autonomous shifts in trading rules. We further introduce a meta-evolutionary outer loop that evolves the prompts guiding program synthesis in the inner loop. This outer loop discovers improved search heuristics. These heuristics balance exploration and exploitation while reducing zero-trade failures. They consistently outperform initial human-designed instructions. The results demonstrate that LLM-based semantic evolution provides a viable approach for continual program synthesis in complex environments.
arXiv:2504.13432v3 Announce Type: replace Abstract: Imaging through inhomogeneous media often results in severe distortions, posing significant challenges to downstream image-processing tasks. The lack of clean paired images makes supervised learning impractical, motivating unsupervised restoration approaches. In this work, we propose the Circular Quasi-Conformal Deturbulence (CQCD) framework, an unsupervised approach that reconstructs distortion-free images from multiple frames using a circular architecture. The framework minimizes reconstruction errors by jointly estimating forward and backward transformations between distorted observations and the restored image. A key advancement of CQCD is the integration of computational quasi-conformal geometry, which encourages bijective non-rigid deformations and improves the well-posedness of both forward and inverse mappings for cycle consistency. The deformation field is further regularized to preserve structural coherence and reduce non-physical artifacts such as folding or tearing. Additionally, tight-frame blocks are employed to effectively encode distortion-sensitive features, enhancing the precision of the restoration process. To assess the effectiveness of the proposed framework, extensive evaluations are conducted on synthetic and real-world image datasets. Experimental findings indicate that CQCD not only surpasses existing state-of-the-art deturbulence techniques in restoration quality but also achieves highly accurate deformation field estimation.
arXiv:2505.00807v2 Announce Type: replace Abstract: Equality saturation, a technique for program optimisation and reasoning, has gained attention due to the resurgence of equality graphs (e-graphs). E-graphs represent equivalence classes of terms under rewrite rules, enabling simultaneous rewriting across a family of terms. However, they struggle in domains like $\lambda$-calculus that involve variable binding, due to a lack of native support for bindings. Building on recent work interpreting e-graphs categorically as morphisms in semilattice-enriched symmetric monoidal categories, we extend this framework to closed symmetric monoidal categories to handle bindings. We provide a concrete combinatorial representation using hierarchical hypergraphs and introduce a corresponding double-pushout (DPO) rewriting mechanism. Finally, we establish the equivalence of term rewriting and DPO rewriting, with the key property that the combinatorial representation absorbs the equations of the symmetric monoidal category.
arXiv:2606.26460v1 Announce Type: new Abstract: AI-based scientific automation is increasingly possible by using agents to generate hypotheses, design experiments, and analyze data. Data collection is a major bottleneck in this pipeline, however. Psychology, and computational cognitive science in particular, is well-positioned to benefit from AI experimentation because theories are often represented as code and crowdsourcing platforms enable programmatic human data collection at scale. Here, we apply automated discovery techniques to the project of generating theories in computational cognitive science, with an agent-based system collecting human data independently through crowdsourced survey experiments. As a testbed, we use a classic case study from cognitive psychology: judging which sequences of coin flips seem subjectively more random. Our system, auto-psych, uses nested agent-based discovery loops to generate explanatory theories of human behavior. The inner loop conjectures, fits, and critiques probabilistic cognitive models; the outer loop designs experiments to test these models, launches them online, and analyzes the data. This system can quickly and reliably recover ground-truth theories from synthetic data via systematic experimentation, but the nested structure is critical to model performance. Further, in three independent sequences of human experiments, the system finds theories that fit the data better than theories generated from the scientific literature. This work thus demonstrates the feasibility of automated data collection and theory discovery in computational cognitive science.
arXiv:2606.26461v1 Announce Type: new Abstract: Label-free vibrational microscopy based on Raman scattering and mid-infrared (MIR) absorption has advanced biological imaging through the development of diverse high-sensitivity modalities. However, rigorous sensitivity comparisons across these techniques remain lacking. Here, we establish a unified theoretical framework for quantitatively comparing the shot-noise-limited sensitivities of stimulated Raman scattering (SRS), MIR photothermal (MIP), and stimulated Raman photothermal (SRP) microscopy. The framework is built on two quantities: an effective absorption coefficient, which places SRS on the same footing as linear absorption, and a photothermal (PT) factor, which isolates the contribution specific to PT readout. Using experimentally realistic parameters, we show that MIR absorption coefficients for representative polar vibrational modes are typically about two orders of magnitude larger than the effective absorption coefficients of SRS, even for strong off-resonant Raman bands. In MIP, however, this intrinsic advantage is partly offset by a small PT factor, limiting the net sensitivity gain over SRS to several-fold and, at most, about one order of magnitude. In SRP, by contrast, burst-pulse thermal accumulation substantially increases the PT factor. Under ideal shot-noise-limited conditions with matched per-pulse excitation-energy conditions in water, SRP is predicted to provide sensitivities several-fold higher than those of SRS and to approach those of MIP. This framework provides a common basis for sensitivity assessment across vibrational imaging modalities and clarifies how absorption strength, photothermal accumulation, and readout conditions determine the practical advantages of each modality.
arXiv:2606.26664v1 Announce Type: new Abstract: Existing homomorphic encryption (HE)-based GNN systems adopt a graph-centric paradigm that couples per-query cost to global graph size, limiting evaluations to at most ~20k nodes and making them incompatible with dynamic, large-scale financial graphs. We propose TGHE (Template-based Graph Homomorphic Encryption), an ego-centric framework that resolves this by exploiting a template phenomenon: local computation trees in transaction graphs converge into a small set of structural shapes. TGHE canonicalizes ego-graphs at the edge and packs structurally identical trees into shared CKKS ciphertexts for SIMD-parallel encrypted inference, with two long-tail optimizers (Approximate Template Fitting and Topology Collapse) ensuring full SIMD coverage. On DGraphFin (3.7M nodes, 4.3M edges), TGHE-Collapse achieves a 66.9x speedup over the sequential encrypted baseline with less than 0.002 AUC loss.