PyTorchKR Papers

최신 논문

총 1,907편
태그·기간 등 상세 필터

카테고리 탐색

로그인 후 이 태그 구독

arXiv:2503.08046공식 코드 · Papers with Code

MultiConIR: Towards multi-condition Information Retrieval

In this paper, we introduce MultiConIR, the first benchmark designed to evaluate retrieval models in multi-condition scenarios. Unlike existing datasets that primarily focus on single-condition queries from search engines, MultiConIR captures real-world complexity by incorporating five diverse domains: books, movies, people, medical cases, and legal documents. We propose three tasks to systematically assess retrieval and reranking models on multi-condition robustness, monotonic relevance ranking, and query format sensitivity. Our findings reveal that existing retrieval and reranking models struggle with multi-condition retrieval, with rerankers suffering severe performance degradation as query complexity increases. We further investigate the performance gap between retrieval and reranking models, exploring potential reasons for these discrepancies, and analysis the impact of different pooling strategies on condition placement sensitivity. Finally, we highlight the strengths of GritLM and Nv-Embed, which demonstrate enhanced adaptability to multi-condition queries, offering insights for future retrieval models. The code and datasets are available at https://github.com/EIT-NLP/MultiConIR.

arXiv:2503.08251

MT-NAM: An Efficient and Adaptive Model for Epileptic Seizure Detection

Enhancing the accuracy and efficiency of machine learning algorithms employed in neural interface systems is crucial for advancing next-generation intelligent therapeutic devices. However, current systems often utilize basic machine learning models that do not fully exploit the natural structure of brain signals. Additionally, existing learning models used for neural signal processing often demonstrate low speed and efficiency during inference. To address these challenges, this study introduces Micro Tree-based NAM (MT-NAM), a distilled model based on the recently proposed Neural Additive Models (NAM). The MT-NAM achieves a remarkable 100$\times$ improvement in inference speed compared to standard NAM, without compromising accuracy. We evaluate our approach on the CHB-MIT scalp EEG dataset, which includes recordings from 24 patients with varying numbers of sessions and seizures. NAM achieves an 85.3\% window-based sensitivity and 95\% specificity. Interestingly, our proposed MT-NAM shows only a 2\% reduction in sensitivity compared to the original NAM. To regain this sensitivity, we utilize a test-time template adjuster (T3A) as an update mechanism, enabling our model to achieve higher sensitivity during test time by accommodating transient shifts in neural signals. With this online update approach, MT-NAM achieves the same sensitivity as the standard NAM while achieving approximately 50$\times$ acceleration in inference speed.

arXiv:2503.07082공식 코드 · Papers with Code

On the Generalization of Representation Uncertainty in Earth Observation

Recent advances in Computer Vision have introduced the concept of pretrained representation uncertainty, enabling zero-shot uncertainty estimation. This holds significant potential for Earth Observation (EO), where trustworthiness is critical, yet the complexity of EO data poses challenges to uncertainty-aware methods. In this work, we investigate the generalization of representation uncertainty in EO, considering the domain's unique semantic characteristics. We pretrain uncertainties on large EO datasets and propose an evaluation framework to assess their zero-shot performance in multi-label classification and segmentation EO tasks. Our findings reveal that, unlike uncertainties pretrained on natural images, EO-pretraining exhibits strong generalization across unseen EO domains, geographic locations, and target granularities, while maintaining sensitivity to variations in ground sampling distance. We demonstrate the practical utility of pretrained uncertainties showcasing their alignment with task-specific uncertainties in downstream tasks, their sensitivity to real-world EO image noise, and their ability to generate spatial uncertainty estimates out-of-the-box. Initiating the discussion on representation uncertainty in EO, our study provides insights into its strengths and limitations, paving the way for future research in the field. Code and weights are available at: https://github.com/Orion-AI-Lab/EOUncertaintyGeneralization.

arXiv:2503.06518

Towards Superior Quantization Accuracy: A Layer-sensitive Approach

Large Vision and Language Models have exhibited remarkable human-like intelligence in tasks such as natural language comprehension, problem-solving, logical reasoning, and knowledge retrieval. However, training and serving these models require substantial computational resources, posing a significant barrier to their widespread application and further research. To mitigate this challenge, various model compression techniques have been developed to reduce computational requirements. Nevertheless, existing methods often employ uniform quantization configurations, failing to account for the varying difficulties across different layers in quantizing large neural network models. This paper tackles this issue by leveraging layer-sensitivity features, such as activation sensitivity and weight distribution Kurtosis, to identify layers that are challenging to quantize accurately and allocate additional memory budget. The proposed methods, named SensiBoost and KurtBoost, respectively, demonstrate notable improvement in quantization accuracy, achieving up to 9% lower perplexity with only a 2% increase in memory budget on LLama models compared to the baseline.

arXiv:2503.07213

Nonlinear Temperature Sensitivity of Residential Electricity Demand: Evidence from a Distributional Regression Approach

We estimate the temperature sensitivity of residential electricity demand during extreme temperature events using the distribution-to-scalar regression model. Rather than relying on simple averages or individual quantile statistics of raw temperature data, we construct distributional summaries, such as probability density, hazard rate, and quantile functions, to retain a more comprehensive representation of temperature variation. This approach not only utilizes richer information from the underlying temperature distribution but also enables the examination of extreme temperature effects that conventional models fail to capture. Additionally, recognizing that distribution functions are typically estimated from limited discrete observations and may be subject to measurement errors, our econometric framework explicitly addresses this issue. Empirical findings from the hazard-to-demand model indicate that residential electricity demand exhibits a stronger nonlinear response to cold waves than to heat waves, while heat wave shocks demonstrate a more pronounced incremental effect. Moreover, the temperature quantile-to-demand model produces largely insignificant demand response estimates, attributed to the offsetting influence of two counteracting forces.

arXiv:2503.04433

Damping Identification Sensitivity in Flutter Speed Estimation

Predicting flutter remains a key challenge in aeroelastic research, with certain models relying on modal parameters, such as natural frequencies and damping ratios. These models are particularly useful in early design stages or for the development of small Unmanned Aerial Vehicles (maximum take-off mass below 7 kg). This study evaluates two frequency-domain system identification methods, Fast Relaxed Vector Fitting (FRVF) and the Loewner Framework (LF), for predicting the flutter onset speed of a flexible wing model. Both methods are applied to extract modal parameters from Ground Vibration Testing data, which are subsequently used to develop a reduced-order model with two degrees of freedom. Results indicate that FRVF and LF-informed models provide reliable flutter speed, with predictions deviating by no more than 3% (FRVF) and 5% (LF) from the N4SID-informed benchmark. The findings highlight the sensitivity of flutter speed predictions to damping ratio identification accuracy and demonstrate the potential of these methods as computationally efficient alternatives for preliminary aeroelastic assessments.

arXiv:2503.05990공식 코드 · Papers with Code

HealthiVert-GAN: A Novel Framework of Pseudo-Healthy Vertebral Image Synthesis for Interpretable Compression Fracture Grading

Osteoporotic vertebral compression fractures (VCFs) are prevalent in the elderly population, typically assessed on computed tomography (CT) scans by evaluating vertebral height loss. This assessment helps determine the fracture's impact on spinal stability and the need for surgical intervention. However, clinical data indicate that many VCFs exhibit irregular compression, complicating accurate diagnosis. While deep learning methods have shown promise in aiding VCFs screening, they often lack interpretability and sufficient sensitivity, limiting their clinical applicability. To address these challenges, we introduce a novel vertebra synthesis-height loss quantification-VCFs grading framework. Our proposed model, HealthiVert-GAN, utilizes a coarse-to-fine synthesis network designed to generate pseudo-healthy vertebral images that simulate the pre-fracture state of fractured vertebrae. This model integrates three auxiliary modules that leverage the morphology and height information of adjacent healthy vertebrae to ensure anatomical consistency. Additionally, we introduce the Relative Height Loss of Vertebrae (RHLV) as a quantification metric, which divides each vertebra into three sections to measure height loss between pre-fracture and post-fracture states, followed by fracture severity classification using a Support Vector Machine (SVM). Our approach achieves state-of-the-art classification performance on both the Verse2019 dataset and our private dataset, and it provides cross-sectional distribution maps of vertebral height loss. This practical tool enhances diagnostic sensitivity in clinical settings and assisting in surgical decision-making. Our code is available: https://github.com/zhibaishouheilab/HealthiVert-GAN.

arXiv:2503.07414

Cost-Effective Design of Grid-tied Community Microgrid

This study aims to develop a cost-effective microgrid design that optimally balances the economic feasibility, reliability, efficiency, and environmental impact in a grid-tied community microgrid. A multi-objective optimization framework is employed, integrating HOMER Pro for system sizing with deep reinforcement learning (DRL). Sensitivity analyses are conducted to evaluate the system performance under varying load demand and renewable energy fluctuations, while an economic sensitivity assessment examines the impact of electricity prices and capital costs on the Levelized Cost of Energy (LCOE). The proposed microgrid configuration achieves high reliability, satisfying 100% of the load, even under adverse weather conditions. The proposed framework attains an efficiency of 91.99% while maintaining a carbon footprint of 302,747 kg/year, which is approximately 95% lower than that of the grid system. The economic analysis indicates a net present cost (NPC) of $4.83M with a competitive LCOE of $0.208/kWh. In addition, the operation cost is $201,473 per year with a capital investment of $1.42M, rendering it a financially viable alternative to conventional grid-dependent systems.This work can be valuable in identifying effective solutions for supplying reliable and cost-effective power to regional and remote areas.

arXiv:2503.00048

Dynamical analysis of an HIV infection model including quiescent cells and immune response

This research gives a thorough examination of an HIV infection model that includes quiescent cells and immune response dynamics in the host. The model, represented by a system of ordinary differential equations, captures the complex interaction between the host's immune response and viral infection. The study focuses on the model's fundamental aspects, such as equilibrium analysis, computing the basic reproduction number $\mathcal{R}_0$, stability analysis, bifurcation phenomena, numerical simulations, and sensitivity analysis. The analysis reveals both an infection equilibrium, which indicates the persistence of the illness, and an infection-free equilibrium, which represents disease control possibilities. Applying matrix-theoretical approaches, stability analysis proved that the infection-free equilibrium is both locally and globally stable for $\mathcal{R}_0 < 1$. For the situation of $\mathcal{R}_0 > 1$, the infection equilibrium is locally asymptotically stable via the Routh--Hurwitz criterion. We also studied the uniform persistence of the infection, demonstrating that the infection remains present above a positive threshold under certain conditions. The study also found a transcritical forward-type bifurcation at $\mathcal{R}_0 = 1$, indicating a critical threshold that affects the system's behavior. The model's temporal dynamics are studied using numerical simulations, and sensitivity analysis identifies the most significant variables by assessing the effects of parameter changes on system behavior.

arXiv:2503.03323

Exchange Rate Sensitivity in Free Zone Trade: An Empirical Study of the Istanbul Ataturk Airport Free Zone

This study as part of an ongoing research effort, empirically examines the relationship between foreign trade in the Istanbul Ataturk Airport Free Zone and exchange rate movements. Monthly data from 2003 to 2016 were analyzed through stationarity tests (Unit Root), followed by the Vector Autoregressive (VAR) model, Cointegration Analysis, and the Toda-Yamamoto Causality Test. The findings indicate that the exchange rate does not significantly affect imports and exports in the free zone. This result suggests that free zones, due to their structural characteristics and operational framework, may be relatively insulated from exchange rate fluctuations. The study contributes to the literature by providing a focused analysis of a specific free zone in Turkiye, highlighting the potential independence of free zone trade from exchange rate volatility.

arXiv:2503.03283공식 코드 · Papers with Code

Exploring specialization and sensitivity of convolutional neural networks in the context of simultaneous image augmentations

Drawing parallels with the way biological networks are studied, we adapt the treatment--control paradigm to explainable artificial intelligence research and enrich it through multi-parametric input alterations. In this study, we propose a framework for investigating the internal inference impacted by input data augmentations. The internal changes in network operation are reflected in activation changes measured by variance, which can be decomposed into components related to each augmentation, employing Sobol indices and Shapley values. These quantities enable one to visualize sensitivity to different variables and use them for guided masking of activations. In addition, we introduce a way of single-class sensitivity analysis where the candidates are filtered according to their matching to prediction bias generated by targeted damaging of the activations. Relying on the observed parallels, we assume that the developed framework can potentially be transferred to studying biological neural networks in complex environments.

arXiv:2503.02858

Efficient Sampling and Sensitivity Analysis of Rare Transient Instability Events via Subset Simulation

Assessing the risk of low-probability high-impact transient instability (TI) events is crucial for ensuring robust and stable power system operation under high uncertainty. However, direct Monte Carlo (DMC) simulation for rare TI event sampling is computationally intensive. This paper proposes a subset simulation-based method for efficient small TI probability estimation, rare TI events sampling, and subsequent sensitivity analysis. Numerical studies on the modified WSCC 9-bus system demonstrate the efficiency of the proposed method over DMC. Additionally, targeted stability enhancement strategies are designed to eliminate rare TI events and enhance the system's robustness to specific transient faults.

arXiv:2503.01622

DOVE: A Large-Scale Multi-Dimensional Predictions Dataset Towards Meaningful LLM Evaluation

Recent work found that LLMs are sensitive to a wide range of arbitrary prompt dimensions, including the type of delimiters, answer enumerators, instruction wording, and more. This throws into question popular single-prompt evaluation practices. We present DOVE (Dataset Of Variation Evaluation) a large-scale dataset containing prompt perturbations of various evaluation benchmarks. In contrast to previous work, we examine LLM sensitivity from an holistic perspective, and assess the joint effects of perturbations along various dimensions, resulting in thousands of perturbations per instance. We evaluate several model families against DOVE, leading to several findings, including efficient methods for choosing well-performing prompts, observing that few-shot examples reduce sensitivity, and identifying instances which are inherently hard across all perturbations. DOVE consists of more than 250M prompt perturbations and model outputs, which we make publicly available to spur a community-wide effort toward meaningful, robust, and efficient evaluation. Browse the data, contribute, and more: https://slab-nlp.github.io/DOVE/

arXiv:2503.00320

Shifting Power: Leveraging LLMs to Simulate Human Aversion in ABMs of Bilateral Financial Exchanges, A bond market study

Bilateral markets, such as those for government bonds, involve decentralized and opaque transactions between market makers (MMs) and clients, posing significant challenges for traditional modeling approaches. To address these complexities, we introduce TRIBE an agent-based model augmented with a large language model (LLM) to simulate human-like decision-making in trading environments. TRIBE leverages publicly available data and stylized facts to capture realistic trading dynamics, integrating human biases like risk aversion and ambiguity sensitivity into the decision-making processes of agents. Our research yields three key contributions: first, we demonstrate that integrating LLMs into agent-based models to enhance client agency is feasible and enriches the simulation of agent behaviors in complex markets; second, we find that even slight trade aversion encoded within the LLM leads to a complete cessation of trading activity, highlighting the sensitivity of market dynamics to agents' risk profiles; third, we show that incorporating human-like variability shifts power dynamics towards clients and can disproportionately affect the entire system, often resulting in systemic agent collapse across simulations. These findings underscore the emergent properties that arise when introducing stochastic, human-like decision processes, revealing new system behaviors that enhance the realism and complexity of artificial societies.

arXiv:2503.01273공식 코드 · Papers with Code

OptMetaOpenFOAM: Large Language Model Driven Chain of Thought for Sensitivity Analysis and Parameter Optimization based on CFD

Merging natural language interfaces with computational fluid dynamics (CFD) workflows presents transformative opportunities for both industry and research. In this study, we introduce OptMetaOpenFOAM - a novel framework that bridges MetaOpenFOAM with external analysis and optimization tool libraries through a large language model (LLM)-driven chain-of-thought (COT) methodology. By automating complex CFD tasks via natural language inputs, the framework empowers non-expert users to perform sensitivity analyses and parameter optimizations with markedly improved efficiency. The test dataset comprises 11 distinct CFD analysis or optimization tasks, including a baseline simulation task derived from an OpenFOAM tutorial covering fluid dynamics, combustion, and heat transfer. Results confirm that OptMetaOpenFOAM can accurately interpret user requirements expressed in natural language and effectively invoke external tool libraries alongside MetaOpenFOAM to complete the tasks. Furthermore, validation on a non-OpenFOAM tutorial case - namely, a hydrogen combustion chamber - demonstrates that a mere 200-character natural language input can trigger a sequence of simulation, postprocessing, analysis, and optimization tasks spanning over 2,000 lines of code. These findings underscore the transformative potential of LLM-driven COT methodologies in linking external tool for advanced analysis and optimization, positioning OptMetaOpenFOAM as an effective tool that streamlines CFD simulations and enhances their convenience and efficiency for both industrial and research applications. Code is available at https://github.com/Terry-cyx/MetaOpenFOAM.

arXiv:2502.20753

Deviance Detection and Regularity Sensitivity in Dissociated Neuronal Cultures

Understanding how neural networks process complex patterns of information is crucial for advancing both neuroscience and artificial intelligence. To investigate fundamental principles of neural computation, we studied dissociated neuronal cultures, one of the most primitive living neural networks, on high-resolution CMOS microelectrode arrays and tested whether the dissociated culture exhibits regularity sensitivity beyond mere stimulus-specific adaptation and deviance detection. In oddball electrical stimulation paradigms, we confirmed that the neuronal culture produced mismatch responses (MMRs) with true deviance detection beyond mere adaptation. These MMRs were dependent on the N-methyl-D-aspartate (NMDA) receptors, similar to mismatch negativity (MMN) in humans, which is known to have true deviance detection properties. Crucially, we also showed sensitivity to the statistical regularity of stimuli, a phenomenon previously observed only in intact brains: the MMRs in a predictable, periodic sequence were smaller than those in a commonly used sequence in which the appearance of the deviant stimulus was random and unpredictable. These results challenge the traditional view that a hierarchically structured neural network is required to process complex temporal patterns, suggesting instead that deviant detection and regularity sensitivity are inherent properties arising from the primitive neural network. They also suggest new directions for the development of neuro-inspired artificial intelligence systems, emphasizing the importance of incorporating adaptive mechanisms and temporal dynamics in the design of neural networks.

arXiv:2502.20952

Efficient Jailbreaking of Large Models by Freeze Training: Lower Layers Exhibit Greater Sensitivity to Harmful Content

With the widespread application of Large Language Models across various domains, their security issues have increasingly garnered significant attention from both academic and industrial communities. This study conducts sampling and normalization of the parameters of the LLM to generate visual representations and heatmaps of parameter distributions, revealing notable discrepancies in parameter distributions among certain layers within the hidden layers. Further analysis involves calculating statistical metrics for each layer, followed by the computation of a Comprehensive Sensitivity Score based on these metrics, which identifies the lower layers as being particularly sensitive to the generation of harmful content. Based on this finding, we employ a Freeze training strategy, selectively performing Supervised Fine-Tuning only on the lower layers. Experimental results demonstrate that this method significantly reduces training duration and GPU memory consumption while maintaining a high jailbreak success rate and a high harm score, outperforming the results achieved by applying the LoRA method for SFT across all layers. Additionally, the method has been successfully extended to other open-source large models, validating its generality and effectiveness across different model architectures. Furthermore, we compare our method with ohter jailbreak method, demonstrating the superior performance of our approach. By innovatively proposing a method to statistically analyze and compare large model parameters layer by layer, this study provides new insights into the interpretability of large models. These discoveries emphasize the necessity of continuous research and the implementation of adaptive security measures in the rapidly evolving field of LLMs to prevent potential jailbreak attack risks, thereby promoting the development of more robust and secure LLMs.

arXiv:2502.10277공식 코드 · Papers with Code

Artificial Intelligence to Assess Dental Findings from Panoramic Radiographs -- A Multinational Study

Dental panoramic radiographs (DPRs) are widely used in clinical practice for comprehensive oral assessment but present challenges due to overlapping structures and time constraints in interpretation. This study aimed to establish a solid baseline for the AI-automated assessment of findings in DPRs by developing, evaluating an AI system, and comparing its performance with that of human readers across multinational data sets. We analyzed 6,669 DPRs from three data sets (the Netherlands, Brazil, and Taiwan), focusing on 8 types of dental findings. The AI system combined object detection and semantic segmentation techniques for per-tooth finding identification. Performance metrics included sensitivity, specificity, and area under the receiver operating characteristic curve (AUC-ROC). AI generalizability was tested across data sets, and performance was compared with human dental practitioners. The AI system demonstrated comparable or superior performance to human readers, particularly +67.9% (95% CI: 54.0%-81.9%; p < .001) sensitivity for identifying periapical radiolucencies and +4.7% (95% CI: 1.4%-8.0%; p = .008) sensitivity for identifying missing teeth. The AI achieved a macro-averaged AUC-ROC of 96.2% (95% CI: 94.6%-97.8%) across 8 findings. AI agreements with the reference were comparable to inter-human agreements in 7 of 8 findings except for caries (p = .024). The AI system demonstrated robust generalization across diverse imaging and demographic settings and processed images 79 times faster (95% CI: 75-82) than human readers. The AI system effectively assessed findings in DPRs, achieving performance on par with or better than human experts while significantly reducing interpretation time. These results highlight the potential for integrating AI into clinical workflows to improve diagnostic efficiency and accuracy, and patient management.

arXiv:2502.10200

Dynamic Reinforcement Learning for Actors

Dynamic Reinforcement Learning (Dynamic RL), proposed in this paper, directly controls system dynamics, instead of the actor (action-generating neural network) outputs at each moment, bringing about a major qualitative shift in reinforcement learning (RL) from static to dynamic. The actor is initially designed to generate chaotic dynamics through the loop with its environment, enabling the agent to perform flexible and deterministic exploration. Dynamic RL controls global system dynamics using a local index called "sensitivity," which indicates how much the input neighborhood contracts or expands into the corresponding output neighborhood through each neuron's processing. While sensitivity adjustment learning (SAL) prevents excessive convergence of the dynamics, sensitivity-controlled reinforcement learning (SRL) adjusts them -- to converge more to improve reproducibility around better state transitions with positive TD error and to diverge more to enhance exploration around worse transitions with negative TD error. Dynamic RL was applied only to the actor in an Actor-Critic RL architecture while applying it to the critic remains a challenge. It was tested on two dynamic tasks and functioned effectively without external exploration noise or backward computation through time. Moreover, it exhibited excellent adaptability to new environments, although some problems remain. Drawing parallels between 'exploration' and 'thinking,' the author hypothesizes that "exploration grows into thinking through learning" and believes this RL could be a key technique for the emergence of thinking, including inspiration that cannot be reconstructed from massive existing text data. Finally, despite being presumptuous, the author presents the argument that this research should not proceed due to its potentially fatal risks, aiming to encourage discussion.

arXiv:2502.20144

Robust sensitivity control in digital pathology via tile score distribution matching

Deploying digital pathology models across medical centers is challenging due to distribution shifts. Recent advances in domain generalization improve model transferability in terms of aggregated performance measured by the Area Under Curve (AUC). However, clinical regulations often require to control the transferability of other metrics, such as prescribed sensitivity levels. We introduce a novel approach to control the sensitivity of whole slide image (WSI) classification models, based on optimal transport and Multiple Instance Learning (MIL). Validated across multiple cohorts and tasks, our method enables robust sensitivity control with only a handful of calibration samples, providing a practical solution for reliable deployment of computational pathology systems.