id
stringlengths
9
16
title
stringlengths
4
278
categories
stringlengths
5
104
abstract
stringlengths
6
4.09k
2501.14646
SyncAnimation: A Real-Time End-to-End Framework for Audio-Driven Human Pose and Talking Head Animation
cs.CV
Generating talking avatar driven by audio remains a significant challenge. Existing methods typically require high computational costs and often lack sufficient facial detail and realism, making them unsuitable for applications that demand high real-time performance and visual quality. Additionally, while some methods can synchronize lip movement, they still face issues with consistency between facial expressions and upper body movement, particularly during silent periods. In this paper, we introduce SyncAnimation, the first NeRF-based method that achieves audio-driven, stable, and real-time generation of speaking avatar by combining generalized audio-to-pose matching and audio-to-expression synchronization. By integrating AudioPose Syncer and AudioEmotion Syncer, SyncAnimation achieves high-precision poses and expression generation, progressively producing audio-synchronized upper body, head, and lip shapes. Furthermore, the High-Synchronization Human Renderer ensures seamless integration of the head and upper body, and achieves audio-sync lip. The project page can be found at https://syncanimation.github.io
2501.14649
Investigating the (De)Composition Capabilities of Large Language Models in Natural-to-Formal Language Conversion
cs.CL
To achieve generalized and robust natural-to-formal language conversion (N2F), large language models (LLMs) need to have strong capabilities of decomposition and composition in N2F when faced with an unfamiliar formal language and be able to cope with compositional gaps and counter-intuitive symbolic names. To investigate whether LLMs have this set of basic capabilities in N2F, we propose the DEDC framework. This framework semi-automatically performs sample and task construction, allowing decoupled evaluation of the set of decomposition and composition capabilities of LLMs in N2F. Based on this framework, we evaluate and analyze the most advanced LLMs, and the main findings include that: (1) the LLMs are deficient in both decomposition and composition; (2) the LLMs show a wide coverage of error types that can be attributed to deficiencies in natural language understanding and the learning and use of symbolic systems; (3) compositional gaps and counter-intuitive symbolic names both affect the decomposition and composition of the LLMs. Our work provides a new perspective for investigating the basic capabilities of decomposition and composition of LLMs in N2F. The detailed analysis of deficiencies and attributions can help subsequent improvements of LLMs.
2501.14652
Decoupled SGDA for Games with Intermittent Strategy Communication
cs.LG
We focus on reducing communication overhead in multiplayer games, where frequently exchanging strategies between players is not feasible and players have noisy or outdated strategies of the other players. We introduce Decoupled SGDA, a novel adaptation of Stochastic Gradient Descent Ascent (SGDA). In this approach, players independently update their strategies based on outdated opponent strategies, with periodic synchronization to align strategies. For Strongly-Convex-Strongly-Concave (SCSC) games, we demonstrate that Decoupled SGDA achieves near-optimal communication complexity comparable to the best-known GDA rates. For weakly coupled games where the interaction between players is lower relative to the non-interactive part of the game, Decoupled SGDA significantly reduces communication costs compared to standard SGDA. Our findings extend to multi-player games. To provide insights into the effect of communication frequency and convergence, we extensively study the convergence of Decoupled SGDA for quadratic minimax problems. Lastly, in settings where the noise over the players is imbalanced, Decoupled SGDA significantly outperforms federated minimax methods.
2501.14653
Federated Domain Generalization with Data-free On-server Gradient Matching
cs.LG cs.AI cs.DC cs.MA
Domain Generalization (DG) aims to learn from multiple known source domains a model that can generalize well to unknown target domains. One of the key approaches in DG is training an encoder which generates domain-invariant representations. However, this approach is not applicable in Federated Domain Generalization (FDG), where data from various domains are distributed across different clients. In this paper, we introduce a novel approach, dubbed Federated Learning via On-server Matching Gradient (FedOMG), which can \emph{efficiently leverage domain information from distributed domains}. Specifically, we utilize the local gradients as information about the distributed models to find an invariant gradient direction across all domains through gradient inner product maximization. The advantages are two-fold: 1) FedOMG can aggregate the characteristics of distributed models on the centralized server without incurring any additional communication cost, and 2) FedOMG is orthogonal to many existing FL/FDG methods, allowing for additional performance improvements by being seamlessly integrated with them. Extensive experimental evaluations on various settings to demonstrate the robustness of FedOMG compared to other FL/FDG baselines. Our method outperforms recent SOTA baselines on four FL benchmark datasets (MNIST, EMNIST, CIFAR-10, and CIFAR-100), and three FDG benchmark datasets (PACS, VLCS, and OfficeHome).
2501.14654
MedAgentBench: A Realistic Virtual EHR Environment to Benchmark Medical LLM Agents
cs.LG cs.AI cs.MA
Recent large language models (LLMs) have demonstrated significant advancements, particularly in their ability to serve as agents thereby surpassing their traditional role as chatbots. These agents can leverage their planning and tool utilization capabilities to address tasks specified at a high level. However, a standardized dataset to benchmark the agent capabilities of LLMs in medical applications is currently lacking, making the evaluation of LLMs on complex tasks in interactive healthcare environments challenging. To address this gap, we introduce MedAgentBench, a broad evaluation suite designed to assess the agent capabilities of large language models within medical records contexts. MedAgentBench encompasses 300 patient-specific clinically-derived tasks from 10 categories written by human physicians, realistic profiles of 100 patients with over 700,000 data elements, a FHIR-compliant interactive environment, and an accompanying codebase. The environment uses the standard APIs and communication infrastructure used in modern EMR systems, so it can be easily migrated into live EMR systems. MedAgentBench presents an unsaturated agent-oriented benchmark that current state-of-the-art LLMs exhibit some ability to succeed at. The best model (Claude 3.5 Sonnet v2) achieves a success rate of 69.67%. However, there is still substantial space for improvement which gives the community a next direction to optimize. Furthermore, there is significant variation in performance across task categories. MedAgentBench establishes this and is publicly available at https://github.com/stanfordmlgroup/MedAgentBench , offering a valuable framework for model developers to track progress and drive continuous improvements in the agent capabilities of large language models within the medical domain.
2501.14659
Towards Unified Structured Light Optimization
cs.CV
Structured light (SL) 3D reconstruction captures the precise surface shape of objects, providing high-accuracy 3D data essential for industrial inspection and robotic vision systems. However, current research on optimizing projection patterns in SL 3D reconstruction faces two main limitations: each scene requires separate training of calibration parameters, and optimization is restricted to specific types of SL, which restricts their application range. To tackle these limitations, we present a unified framework for SL optimization, adaptable to diverse lighting conditions, object types, and different types of SL. Our framework quickly determines the optimal projection pattern using only a single projected image. Key contributions include a novel global matching method for projectors, enabling precise projector-camera alignment with just one projected image, and a new projection compensation model with a photometric adjustment module to reduce artifacts from out-of-gamut clipping. Experimental results show our method achieves superior decoding accuracy across various objects, SL patterns, and lighting conditions, significantly outperforming previous methods.
2501.14660
Mean-field limit from general mixtures of experts to quantum neural networks
math-ph cs.LG math.MP math.PR
In this work, we study the asymptotic behavior of Mixture of Experts (MoE) trained via gradient flow on supervised learning problems. Our main result establishes the propagation of chaos for a MoE as the number of experts diverges. We demonstrate that the corresponding empirical measure of their parameters is close to a probability measure that solves a nonlinear continuity equation, and we provide an explicit convergence rate that depends solely on the number of experts. We apply our results to a MoE generated by a quantum neural network.
2501.14661
Neural-Symbolic Message Passing with Dynamic Pruning
cs.LG cs.AI
Complex Query Answering (CQA) over incomplete Knowledge Graphs (KGs) is a challenging task. Recently, a line of message-passing-based research has been proposed to solve CQA. However, they perform unsatisfactorily on negative queries and fail to address the noisy messages between variable nodes in the query graph. Moreover, they offer little interpretability and require complex query data and resource-intensive training. In this paper, we propose a Neural-Symbolic Message Passing (NSMP) framework based on pre-trained neural link predictors. By introducing symbolic reasoning and fuzzy logic, NSMP can generalize to arbitrary existential first order logic queries without requiring training while providing interpretable answers. Furthermore, we introduce a dynamic pruning strategy to filter out noisy messages between variable nodes. Experimental results show that NSMP achieves a strong performance. Additionally, through complexity analysis and empirical verification, we demonstrate the superiority of NSMP in inference time over the current state-of-the-art neural-symbolic method. Compared to this approach, NSMP demonstrates faster inference times across all query types on benchmark datasets, with speedup ranging from 2$\times$ to over 150$\times$.
2501.14663
End-to-end workflow for machine learning-based qubit readout with QICK and hls4ml
quant-ph cs.LG
We present an end-to-end workflow for superconducting qubit readout that embeds co-designed Neural Networks (NNs) into the Quantum Instrumentation Control Kit (QICK). Capitalizing on the custom firmware and software of the QICK platform, which is built on Xilinx RFSoC FPGAs, we aim to leverage machine learning (ML) to address critical challenges in qubit readout accuracy and scalability. The workflow utilizes the hls4ml package and employs quantization-aware training to translate ML models into hardware-efficient FPGA implementations via user-friendly Python APIs. We experimentally demonstrate the design, optimization, and integration of an ML algorithm for single transmon qubit readout, achieving 96% single-shot fidelity with a latency of 32ns and less than 16% FPGA look-up table resource utilization. Our results offer the community an accessible workflow to advance ML-driven readout and adaptive control in quantum information processing applications.
2501.14664
Predictive Position Estimation for Remote Surgery under Packet Loss Using the Informer Framework
eess.SY cs.SY
Accurate and real-time position estimation of the robotic arm on the patient's side is crucial for the success of remote robotic surgery in Tactile Internet environments. This paper proposes a predictive approach using the computationally efficient Transformer-based Informer model for position estimation, combined with a Four-State Hidden Markov Model (4-State HMM) to simulate realistic packet loss scenarios. The method effectively addresses network-induced delays, jitter, and packet loss, ensuring reliable performance in remote robotic surgery. The study evaluates the Informer model on the JIGSAWS dataset, demonstrating its capability to handle sequential data challenges caused by network uncertainties. Key features, including ProbSparse attention and a generative-style decoder, enhance prediction accuracy, computational speed, and memory efficiency. Results indicate that the proposed method achieves over 90 percent accuracy across varying network conditions. Furthermore, the Informer framework outperforms traditional models such as TCN, RNN, and LSTM, highlighting its suitability for real-time remote surgery applications.
2501.14672
Gaussian-Process-based Adaptive Tracking Control with Dynamic Active Learning for Autonomous Ground Vehicles
eess.SY cs.RO cs.SY
This article proposes an active-learning-based adaptive trajectory tracking control method for autonomous ground vehicles to compensate for modeling errors and unmodeled dynamics. The nominal vehicle model is decoupled into lateral and longitudinal subsystems, which are augmented with online Gaussian Processes (GPs), using measurement data. The estimated mean functions of the GPs are used to construct a feedback compensator, which, together with an LPV state feedback controller designed for the nominal system, gives the adaptive control structure. To assist exploration of the dynamics, the paper proposes a new, dynamic active learning method to collect the most informative samples to accelerate the training process. To analyze the performance of the overall learning tool-chain provided controller, a novel iterative, counterexample-based algorithm is proposed for calculating the induced L2 gain between the reference trajectory and the tracking error. The analysis can be executed for a set of possible realizations of the to-be-controlled system, giving robust performance certificate of the learning method under variation of the vehicle dynamics. The efficiency of the proposed control approach is shown on a high-fidelity physics simulator and in real experiments using a 1/10 scale F1TENTH electric car.
2501.14673
State Space Models for Extractive Summarization in Low Resource Scenarios
cs.CL cs.AI
Extractive summarization involves selecting the most relevant sentences from a text. Recently, researchers have focused on advancing methods to improve state-of-the-art results in low-resource settings. Motivated by these advancements, we propose the MPoincareSum method. This method applies the Mamba state space model to generate the semantics of reviews and sentences, which are then concatenated. A Poincare compression is used to select the most meaningful features, followed by the application of a linear layer to predict sentence relevance based on the corresponding review. Finally, we paraphrase the relevant sentences to create the final summary. To evaluate the effectiveness of MPoincareSum, we conducted extensive experiments using the Amazon review dataset. The performance of the method was assessed using ROUGE scores. The experimental results demonstrate that MPoincareSum outperforms several existing approaches in the literature
2501.14677
MatAnyone: Stable Video Matting with Consistent Memory Propagation
cs.CV
Auxiliary-free human video matting methods, which rely solely on input frames, often struggle with complex or ambiguous backgrounds. To address this, we propose MatAnyone, a robust framework tailored for target-assigned video matting. Specifically, building on a memory-based paradigm, we introduce a consistent memory propagation module via region-adaptive memory fusion, which adaptively integrates memory from the previous frame. This ensures semantic stability in core regions while preserving fine-grained details along object boundaries. For robust training, we present a larger, high-quality, and diverse dataset for video matting. Additionally, we incorporate a novel training strategy that efficiently leverages large-scale segmentation data, boosting matting stability. With this new network design, dataset, and training strategy, MatAnyone delivers robust and accurate video matting results in diverse real-world scenarios, outperforming existing methods.
2501.14678
A Predictive Approach for Enhancing Accuracy in Remote Robotic Surgery Using Informer Model
cs.RO cs.AI
Precise and real-time estimation of the robotic arm's position on the patient's side is essential for the success of remote robotic surgery in Tactile Internet (TI) environments. This paper presents a prediction model based on the Transformer-based Informer framework for accurate and efficient position estimation. Additionally, it combines a Four-State Hidden Markov Model (4-State HMM) to simulate realistic packet loss scenarios. The proposed approach addresses challenges such as network delays, jitter, and packet loss to ensure reliable and precise operation in remote surgical applications. The method integrates the optimization problem into the Informer model by embedding constraints such as energy efficiency, smoothness, and robustness into its training process using a differentiable optimization layer. The Informer framework uses features such as ProbSparse attention, attention distilling, and a generative-style decoder to focus on position-critical features while maintaining a low computational complexity of O(L log L). The method is evaluated using the JIGSAWS dataset, achieving a prediction accuracy of over 90 percent under various network scenarios. A comparison with models such as TCN, RNN, and LSTM demonstrates the Informer framework's superior performance in handling position prediction and meeting real-time requirements, making it suitable for Tactile Internet-enabled robotic surgery.
2501.14679
Surface Vision Mamba: Leveraging Bidirectional State Space Model for Efficient Spherical Manifold Representation
cs.CV cs.AI
Attention-based methods have demonstrated exceptional performance in modelling long-range dependencies on spherical cortical surfaces, surpassing traditional Geometric Deep Learning (GDL) models. However, their extensive inference time and high memory demands pose challenges for application to large datasets with limited computing resources. Inspired by the state space model in computer vision, we introduce the attention-free Vision Mamba (Vim) to spherical surfaces, presenting a domain-agnostic architecture for analyzing data on spherical manifolds. Our method achieves surface patching by representing spherical data as a sequence of triangular patches derived from a subdivided icosphere. The proposed Surface Vision Mamba (SiM) is evaluated on multiple neurodevelopmental phenotype regression tasks using cortical surface metrics from neonatal brains. Experimental results demonstrate that SiM outperforms both attention- and GDL-based methods, delivering 4.8 times faster inference and achieving 91.7% lower memory consumption compared to the Surface Vision Transformer (SiT) under the Ico-4 grid partitioning. Sensitivity analysis further underscores the potential of SiM to identify subtle cognitive developmental patterns. The code is available at https://github.com/Rongzhao-He/surface-vision-mamba.
2501.14685
Rethinking Foundation Models for Medical Image Classification through a Benchmark Study on MedMNIST
eess.IV cs.AI cs.CV cs.LG
Foundation models are widely employed in medical image analysis, due to their high adaptability and generalizability for downstream tasks. With the increasing number of foundation models being released, model selection has become an important issue. In this work, we study the capabilities of foundation models in medical image classification tasks by conducting a benchmark study on the MedMNIST dataset. Specifically, we adopt various foundation models ranging from convolutional to Transformer-based models and implement both end-to-end training and linear probing for all classification tasks. The results demonstrate the significant potential of these pre-trained models when transferred for medical image classification. We further conduct experiments with different image sizes and various sizes of training data. By analyzing all the results, we provide preliminary, yet useful insights and conclusions on this topic.
2501.14687
Decoding Generalization from Memorization in Deep Neural Networks
cs.LG cs.AI
Overparameterized Deep Neural Networks that generalize well have been key to the dramatic success of Deep Learning in recent years. The reasons for their remarkable ability to generalize are not well understood yet. It has also been known that deep networks possess the ability to memorize training data, as evidenced by perfect or high training accuracies on models trained with corrupted data that have class labels shuffled to varying degrees. Concomitantly, such models are known to generalize poorly, i.e. they suffer from poor test accuracies, due to which it is thought that the act of memorizing substantially degrades the ability to generalize. It has, however, been unclear why the poor generalization that accompanies such memorization, comes about. One possibility is that in the process of training with corrupted data, the layers of the network irretrievably reorganize their representations in a manner that makes generalization difficult. The other possibility is that the network retains significant ability to generalize, but the trained network somehow chooses to readout in a manner that is detrimental to generalization. Here, we provide evidence for the latter possibility by demonstrating, empirically, that such models possess information in their representations for substantially improved generalization, even in the face of memorization. Furthermore, such generalization abilities can be easily decoded from the internals of the trained model, and we build a technique to do so from the outputs of specific layers of the network. We demonstrate results on multiple models trained with a number of standard datasets.
2501.14689
Approach to Designing CV Systems for Medical Applications: Data, Architecture and AI
cs.CV cs.AI
This paper introduces an innovative software system for fundus image analysis that deliberately diverges from the conventional screening approach, opting not to predict specific diagnoses. Instead, our methodology mimics the diagnostic process by thoroughly analyzing both normal and pathological features of fundus structures, leaving the ultimate decision-making authority in the hands of healthcare professionals. Our initiative addresses the need for objective clinical analysis and seeks to automate and enhance the clinical workflow of fundus image examination. The system, from its overarching architecture to the modular analysis design powered by artificial intelligence (AI) models, aligns seamlessly with ophthalmological practices. Our unique approach utilizes a combination of state-of-the-art deep learning methods and traditional computer vision algorithms to provide a comprehensive and nuanced analysis of fundus structures. We present a distinctive methodology for designing medical applications, using our system as an illustrative example. Comprehensive verification and validation results demonstrate the efficacy of our approach in revolutionizing fundus image analysis, with potential applications across various medical domains.
2501.14693
Rethinking Table Instruction Tuning
cs.CL cs.AI
Recent advances in table understanding have focused on instruction-tuning large language models (LLMs) for table-related tasks. However, existing research has overlooked the impact of hyperparameter choices and lacks a comprehensive evaluation of the out-of-domain table understanding ability and the general capabilities of these table LLMs. In this paper, we evaluate these abilities in existing table LLMs, and reveal significant declines in both out-of-domain table understanding and general capabilities compared to their base models. Through systematic analysis, we show that hyperparameters, such as learning rate, can significantly influence both table-specific and general capabilities. Contrary to the existing table instruction-tuning works, we demonstrate that smaller learning rates and fewer training instances can enhance table understanding while preserving general capabilities. Based on our findings, we introduce TAMA, a TAble LLM instruction-tuned from LLaMA 3.1 8B Instruct, which achieves performance on par with, or surpassing GPT-3.5 and GPT-4 on table tasks, while maintaining strong out-of-domain generalization and general capabilities. Our findings highlight the potential for reduced data annotation costs and more efficient model development through careful hyperparameter selection.
2501.14694
Towards Automated Self-Supervised Learning for Truly Unsupervised Graph Anomaly Detection
cs.LG cs.AI
Self-supervised learning (SSL) is an emerging paradigm that exploits supervisory signals generated from the data itself, and many recent studies have leveraged SSL to conduct graph anomaly detection. However, we empirically found that three important factors can substantially impact detection performance across datasets: 1) the specific SSL strategy employed; 2) the tuning of the strategy's hyperparameters; and 3) the allocation of combination weights when using multiple strategies. Most SSL-based graph anomaly detection methods circumvent these issues by arbitrarily or selectively (i.e., guided by label information) choosing SSL strategies, hyperparameter settings, and combination weights. While an arbitrary choice may lead to subpar performance, using label information in an unsupervised setting is label information leakage and leads to severe overestimation of a method's performance. Leakage has been criticized as "one of the top ten data mining mistakes", yet many recent studies on SSL-based graph anomaly detection have been using label information to select hyperparameters. To mitigate this issue, we propose to use an internal evaluation strategy (with theoretical analysis) to select hyperparameters in SSL for unsupervised anomaly detection. We perform extensive experiments using 10 recent SSL-based graph anomaly detection algorithms on various benchmark datasets, demonstrating both the prior issues with hyperparameter selection and the effectiveness of our proposed strategy.
2501.14696
Predictor-Feedback Stabilization of Globally Lipschitz Nonlinear Systems with State and Input Quantization
math.OC cs.SY eess.SY
We develop a switched nonlinear predictor-feedback control law to achieve global asymptotic stabilization for nonlinear systems with arbitrarily long input delay, under state quantization. The proposed design generalizes the nonlinear predictor-feedback framework by incorporating quantized measurements of both the plant and actuator states into the predictor state formulation. Due to the mismatch between the (inapplicable) exact predictor state and the predictor state constructed in the presence of state quantization, a global stabilization result is possible under a global Lipschitzness assumption on the vector field, as well as under the assumption of existence of a globally Lipschitz, nominal feedback law that achieves global exponential stability of the delay and quantization-free system. To address the constraints imposed by quantization, a dynamic switching strategy is constructed, adjusting the quantizer's tunable parameter in a piecewise constant manner-initially increasing the quantization range, to capture potentially large system states and subsequently refining the precision to reduce quantization error. The global asymptotic stability of the closed-loop system is established through solutions estimates derived using backstepping transformations, combined with small-gain and input-to-state stability arguments. We also extend our approach to the case of input quantization.
2501.14700
An Attentive Graph Agent for Topology-Adaptive Cyber Defence
cs.LG cs.AI cs.CR cs.NI
As cyber threats grow increasingly sophisticated, reinforcement learning (RL) is emerging as a promising technique to create intelligent and adaptive cyber defense systems. However, most existing autonomous defensive agents have overlooked the inherent graph structure of computer networks subject to cyber attacks, potentially missing critical information and constraining their adaptability. To overcome these limitations, we developed a custom version of the Cyber Operations Research Gym (CybORG) environment, encoding network state as a directed graph with realistic low-level features. We employ a Graph Attention Network (GAT) architecture to process node, edge, and global features, and adapt its output to be compatible with policy gradient methods in RL. Our GAT-based approach offers key advantages over flattened alternatives: policies that demonstrate resilience to certain types of unexpected dynamic network topology changes, reasonable generalisation to networks of varying sizes within the same structural distribution, and interpretable defensive actions grounded in tangible network properties. We demonstrate that GAT defensive policies can be trained using our low-level directed graph observations, even when unexpected connections arise during simulation. Evaluations across networks of different sizes, but consistent subnetwork structure, show our policies achieve comparable performance to policies trained specifically for each network configuration. Our study contributes to the development of robust cyber defence systems that can better adapt to real-world network security challenges.
2501.14701
NLP-based assessment of prescription appropriateness from Italian referrals
cs.CL cs.LG
Objective: This study proposes a Natural Language Processing pipeline to evaluate prescription appropriateness in Italian referrals, where reasons for prescriptions are recorded only as free text, complicating automated comparisons with guidelines. The pipeline aims to derive, for the first time, a comprehensive summary of the reasons behind these referrals and a quantification of their appropriateness. While demonstrated in a specific case study, the approach is designed to generalize to other types of examinations. Methods: Leveraging embeddings from a transformer-based model, the proposed approach clusters referral texts, maps clusters to labels, and aligns these labels with existing guidelines. We present a case study on a dataset of 496,971 referrals, consisting of all referrals for venous echocolordopplers of the lower limbs between 2019 and 2021 in the Lombardy Region. A sample of 1,000 referrals was manually annotated to validate the results. Results: The pipeline exhibited high performance for referrals' reasons (Prec=92.43%, Rec=83.28%) and excellent results for referrals' appropriateness (Prec=93.58%, Rec=91.52%) on the annotated subset. Analysis of the entire dataset identified clusters matching guideline-defined reasons - both appropriate and inappropriate - as well as clusters not addressed in the guidelines. Overall, 34.32% of referrals were marked as appropriate, 34.07% inappropriate, 14.37% likely inappropriate, and 17.24% could not be mapped to guidelines. Conclusions: The proposed pipeline effectively assessed prescription appropriateness across a large dataset, serving as a valuable tool for health authorities. Findings have informed the Lombardy Region's efforts to strengthen recommendations and reduce the burden of inappropriate referrals.
2501.14704
Stroke classification using Virtual Hybrid Edge Detection from in silico electrical impedance tomography data
math.AP cs.CV cs.NA math.NA
Electrical impedance tomography (EIT) is a non-invasive imaging method for recovering the internal conductivity of a physical body from electric boundary measurements. EIT combined with machine learning has shown promise for the classification of strokes. However, most previous works have used raw EIT voltage data as network inputs. We build upon a recent development which suggested the use of special noise-robust Virtual Hybrid Edge Detection (VHED) functions as network inputs, although that work used only highly simplified and mathematically ideal models. In this work we strengthen the case for the use of EIT, and VHED functions especially, for stroke classification. We design models with high detail and mathematical realism to test the use of VHED functions as inputs. Virtual patients are created using a physically detailed 2D head model which includes features known to create challenges in real-world imaging scenarios. Conductivity values are drawn from statistically realistic distributions, and phantoms are afflicted with either hemorrhagic or ischemic strokes of various shapes and sizes. Simulated noisy EIT electrode data, generated using the realistic Complete Electrode Model (CEM) as opposed to the mathematically ideal continuum model, is processed to obtain VHED functions. We compare the use of VHED functions as inputs against the alternative paradigm of using raw EIT voltages. Our results show that (i) stroke classification can be performed with high accuracy using 2D EIT data from physically detailed and mathematically realistic models, and (ii) in the presence of noise, VHED functions outperform raw data as network inputs.
2501.14705
The Karp Dataset
cs.LG cs.CL
Understanding the mathematical reasoning capabilities of Large Language Models (LLMs) is a central topic in the study of artificial intelligence. This new domain necessitates the creation of datasets of reasoning tasks for both training and benchmarking the performance of LLMs. To this end, we introduce the Karp dataset: The first dataset composed of detailed proofs of NP-completeness reductions. The reductions vary in difficulty, ranging from simple exercises of undergraduate courses to more challenging reductions from academic papers. We compare the performance of state-of-the-art models on this task and demonstrate the effect of fine-tuning with the Karp dataset on reasoning capacity.
2501.14708
Decision-Focused Learning for Complex System Identification: HVAC Management System Application
eess.SY cs.LG cs.SY
As opposed to conventional training methods tailored to minimize a given statistical metric or task-agnostic loss (e.g., mean squared error), Decision-Focused Learning (DFL) trains machine learning models for optimal performance in downstream decision-making tools. We argue that DFL can be leveraged to learn the parameters of system dynamics, expressed as constraint of the convex optimization control policy, while the system control signal is being optimized, thus creating an end-to-end learning framework. This is particularly relevant for systems in which behavior changes once the control policy is applied, hence rendering historical data less applicable. The proposed approach can perform system identification - i.e., determine appropriate parameters for the system analytical model - and control simultaneously to ensure that the model's accuracy is focused on areas most relevant to control. Furthermore, because black-box systems are non-differentiable, we design a loss function that requires solely to measure the system response. We propose pre-training on historical data and constraint relaxation to stabilize the DFL and deal with potential infeasibilities in learning. We demonstrate the usefulness of the method on a building Heating, Ventilation, and Air Conditioning day-ahead management system for a realistic 15-zone building located in Denver, US. The results show that the conventional RC building model, with the parameters obtained from historical data using supervised learning, underestimates HVAC electrical power consumption. For our case study, the ex-post cost is on average six times higher than the expected one. Meanwhile, the same RC model with parameters obtained via DFL underestimates the ex-post cost only by 3%.
2501.14709
Enhanced Confocal Laser Scanning Microscopy with Adaptive Physics Informed Deep Autoencoders
cond-mat.mtrl-sci cs.CV eess.IV
We present a physics-informed deep learning framework to address common limitations in Confocal Laser Scanning Microscopy (CLSM), such as diffraction limited resolution, noise, and undersampling due to low laser power conditions. The optical system's point spread function (PSF) and common CLSM image degradation mechanisms namely photon shot noise, dark current noise, motion blur, speckle noise, and undersampling were modeled and were directly included into model architecture. The model reconstructs high fidelity images from heavily noisy inputs by using convolutional and transposed convolutional layers. Following the advances in compressed sensing, our approach significantly reduces data acquisition requirements without compromising image resolution. The proposed method was extensively evaluated on simulated CLSM images of diverse structures, including lipid droplets, neuronal networks, and fibrillar systems. Comparisons with traditional deconvolution algorithms such as Richardson-Lucy (RL), non-negative least squares (NNLS), and other methods like Total Variation (TV) regularization, Wiener filtering, and Wavelet denoising demonstrate the superiority of the network in restoring fine structural details with high fidelity. Assessment metrics like Structural Similarity Index (SSIM) and Peak Signal to Noise Ratio (PSNR), underlines that the AdaptivePhysicsAutoencoder achieved robust image enhancement across diverse CLSM conditions, helping faster acquisition, reduced photodamage, and reliable performance in low light and sparse sampling scenarios holding promise for applications in live cell imaging, dynamic biological studies, and high throughput material characterization.
2501.14710
Overcoming Fairness Trade-offs via Pre-processing: A Causal Perspective
stat.ML cs.LG
Training machine learning models for fair decisions faces two key challenges: The \emph{fairness-accuracy trade-off} results from enforcing fairness which weakens its predictive performance in contrast to an unconstrained model. The incompatibility of different fairness metrics poses another trade-off -- also known as the \emph{impossibility theorem}. Recent work identifies the bias within the observed data as a possible root cause and shows that fairness and predictive performance are in fact in accord when predictive performance is measured on unbiased data. We offer a causal explanation for these findings using the framework of the FiND (fictitious and normatively desired) world, a "fair" world, where protected attributes have no causal effects on the target variable. We show theoretically that (i) classical fairness metrics deemed to be incompatible are naturally satisfied in the FiND world, while (ii) fairness aligns with high predictive performance. We extend our analysis by suggesting how one can benefit from these theoretical insights in practice, using causal pre-processing methods that approximate the FiND world. Additionally, we propose a method for evaluating the approximation of the FiND world via pre-processing in practical use cases where we do not have access to the FiND world. In simulations and empirical studies, we demonstrate that these pre-processing methods are successful in approximating the FiND world and resolve both trade-offs. Our results provide actionable solutions for practitioners to achieve fairness and high predictive performance simultaneously.
2501.14713
FlexiGPT: Pruning and Extending Large Language Models with Low-Rank Weight Sharing
cs.CL cs.LG
The rapid proliferation of large language models (LLMs) in natural language processing (NLP) has created a critical need for techniques that enable efficient deployment on memory-constrained devices without compromising performance. We present a method to prune LLMs that selectively prunes model blocks based on an importance score and replaces them with a low-parameter replacement strategy. Specifically, we propose a principled metric to replace each pruned block using a weight-sharing mechanism that leverages unpruned counterparts from the model and block-specific low-rank adapters. Furthermore, we facilitate the learning of these replacement blocks with output feature normalization and an adapter initialization scheme built on low-rank SVD reconstructions. Empirical evaluations demonstrate substantial performance gains over existing methods, achieving state-of-the-art performance on 5/6 benchmarks for a compression rate of 30% and 6/6 benchmarks for a compression rate of 40%. We also demonstrate that our approach can extend smaller models, boosting performance on 6/6 benchmarks using only ~0.3% tokens of extended training with minimal additional parameter costs.
2501.14717
Towards Better Understanding Table Instruction Tuning: Decoupling the Effects from Data versus Models
cs.CL
Recent advances in natural language processing have leveraged instruction tuning to enhance Large Language Models (LLMs) for table-related tasks. However, previous works train different base models with different training data, lacking an apples-to-apples comparison across the result table LLMs. To address this, we fine-tune base models from the Mistral, OLMo, and Phi families on existing public training datasets. Our replication achieves performance on par with or surpassing existing table LLMs, establishing new state-of-the-art performance on Hitab, a table question-answering dataset. More importantly, through systematic out-of-domain evaluation, we decouple the contributions of training data and the base model, providing insight into their individual impacts. In addition, we assess the effects of table-specific instruction tuning on general-purpose benchmarks, revealing trade-offs between specialization and generalization.
2501.14719
Do LLMs Provide Consistent Answers to Health-Related Questions across Languages?
cs.CL cs.AI cs.HC cs.IR
Equitable access to reliable health information is vital for public health, but the quality of online health resources varies by language, raising concerns about inconsistencies in Large Language Models (LLMs) for healthcare. In this study, we examine the consistency of responses provided by LLMs to health-related questions across English, German, Turkish, and Chinese. We largely expand the HealthFC dataset by categorizing health-related questions by disease type and broadening its multilingual scope with Turkish and Chinese translations. We reveal significant inconsistencies in responses that could spread healthcare misinformation. Our main contributions are 1) a multilingual health-related inquiry dataset with meta-information on disease categories, and 2) a novel prompt-based evaluation workflow that enables sub-dimensional comparisons between two languages through parsing. Our findings highlight key challenges in deploying LLM-based tools in multilingual contexts and emphasize the need for improved cross-lingual alignment to ensure accurate and equitable healthcare information.
2501.14720
Communication-Based Distributed Control of Large-Scale District Heating Networks
eess.SY cs.SY
This paper presents a non-cooperative distributed model predictive controller for the control of large-scale District Heating Networks. To enable the design of this controller a novel information passing scheme and feasibility restoration method are created, allowing the local controllers to achieve a global consensus while minimizing a local cost function. The effectiveness of this controller is demonstrated on an 18-user District Heating Network decomposed into six subsystems. The results show that the developed control scheme effectively uses flexibility to manage the buildings' heat demands reducing the total losses by 14% and the return temperature by 37%.
2501.14721
Comparable Corpora: Opportunities for New Research Directions
cs.CL
Most conference papers present new results, but this paper will focus more on opportunities for the audience to make their own contributions. This paper is intended to challenge the community to think more broadly about what we can do with comparable corpora. We will start with a review of the history, and then suggest new directions for future research. This was a keynote at BUCC-2025, a workshop associated with Coling-2025.
2501.14723
CodeMonkeys: Scaling Test-Time Compute for Software Engineering
cs.LG
Scaling test-time compute is a promising axis for improving LLM capabilities. However, test-time compute can be scaled in a variety of ways, and effectively combining different approaches remains an active area of research. Here, we explore this problem in the context of solving real-world GitHub issues from the SWE-bench dataset. Our system, named CodeMonkeys, allows models to iteratively edit a codebase by jointly generating and running a testing script alongside their draft edit. We sample many of these multi-turn trajectories for every issue to generate a collection of candidate edits. This approach lets us scale "serial" test-time compute by increasing the number of iterations per trajectory and "parallel" test-time compute by increasing the number of trajectories per problem. With parallel scaling, we can amortize up-front costs across multiple downstream samples, allowing us to identify relevant codebase context using the simple method of letting an LLM read every file. In order to select between candidate edits, we combine voting using model-generated tests with a final multi-turn trajectory dedicated to selection. Overall, CodeMonkeys resolves 57.4% of issues from SWE-bench Verified using a budget of approximately 2300 USD. Our selection method can also be used to combine candidates from different sources. Selecting over an ensemble of edits from existing top SWE-bench Verified submissions obtains a score of 66.2% and outperforms the best member of the ensemble on its own. We fully release our code and data at https://scalingintelligence.stanford.edu/pubs/codemonkeys.
2501.14724
MLPs at the EOC: Concentration of the NTK
cs.LG stat.ML
We study the concentration of the Neural Tangent Kernel (NTK) $K_\theta : \mathbb{R}^{m_0} \times \mathbb{R}^{m_0} \to \mathbb{R}^{m_l \times m_l}$ of $l$-layer Multilayer Perceptrons (MLPs) $N : \mathbb{R}^{m_0} \times \Theta \to \mathbb{R}^{m_l}$ equipped with activation functions $\phi(s) = a s + b \vert s \vert$ for some $a,b \in \mathbb{R}$ with the parameter $\theta \in \Theta$ being initialized at the Edge Of Chaos (EOC). Without relying on the gradient independence assumption that has only been shown to hold asymptotically in the infinitely wide limit, we prove that an approximate version of gradient independence holds at finite width. Showing that the NTK entries $K_\theta(x_{i_1},x_{i_2})$ for $i_1,i_2 \in [1:n]$ over a dataset $\{x_1,\cdots,x_n\} \subset \mathbb{R}^{m_0}$ concentrate simultaneously via maximal inequalities, we prove that the NTK matrix $K(\theta) = [\frac{1}{n} K_\theta(x_{i_1},x_{i_2}) : i_1,i_2 \in [1:n]] \in \mathbb{R}^{nm_l \times nm_l}$ concentrates around its infinitely wide limit $\overset{\scriptscriptstyle\infty}{K} \in \mathbb{R}^{nm_l \times nm_l}$ without the need for linear overparameterization. Our results imply that in order to accurately approximate the limit, hidden layer widths have to grow quadratically as $m_k = k^2 m$ for some $m \in \mathbb{N}+1$ for sufficient concentration. For such MLPs, we obtain the concentration bound $\mathbb{P}( \Vert K(\theta) - \overset{\scriptscriptstyle\infty}{K} \Vert \leq O((\Delta_\phi^{-2} + m_l^{\frac{1}{2}} l) \kappa_\phi^2 m^{-\frac{1}{2}})) \geq 1-O(m^{-1})$ modulo logarithmic terms, where we denoted $\Delta_\phi = \frac{b^2}{a^2+b^2}$ and $\kappa_\phi = \frac{\vert a \vert + \vert b \vert}{\sqrt{a^2 + b^2}}$. This reveals in particular that the absolute value ($\Delta_\phi=1$, $\kappa_\phi=1$) beats the ReLU ($\Delta_\phi=\frac{1}{2}$, $\kappa_\phi=\sqrt{2}$) in terms of the concentration of the NTK.
2501.14726
Relightable Full-Body Gaussian Codec Avatars
cs.CV cs.GR
We propose Relightable Full-Body Gaussian Codec Avatars, a new approach for modeling relightable full-body avatars with fine-grained details including face and hands. The unique challenge for relighting full-body avatars lies in the large deformations caused by body articulation and the resulting impact on appearance caused by light transport. Changes in body pose can dramatically change the orientation of body surfaces with respect to lights, resulting in both local appearance changes due to changes in local light transport functions, as well as non-local changes due to occlusion between body parts. To address this, we decompose the light transport into local and non-local effects. Local appearance changes are modeled using learnable zonal harmonics for diffuse radiance transfer. Unlike spherical harmonics, zonal harmonics are highly efficient to rotate under articulation. This allows us to learn diffuse radiance transfer in a local coordinate frame, which disentangles the local radiance transfer from the articulation of the body. To account for non-local appearance changes, we introduce a shadow network that predicts shadows given precomputed incoming irradiance on a base mesh. This facilitates the learning of non-local shadowing between the body parts. Finally, we use a deferred shading approach to model specular radiance transfer and better capture reflections and highlights such as eye glints. We demonstrate that our approach successfully models both the local and non-local light transport required for relightable full-body avatars, with a superior generalization ability under novel illumination conditions and unseen poses.
2501.14728
Mitigating GenAI-powered Evidence Pollution for Out-of-Context Multimodal Misinformation Detection
cs.MM cs.CL cs.CV cs.CY
While large generative artificial intelligence (GenAI) models have achieved significant success, they also raise growing concerns about online information security due to their potential misuse for generating deceptive content. Out-of-context (OOC) multimodal misinformation detection, which often retrieves Web evidence to identify the repurposing of images in false contexts, faces the issue of reasoning over GenAI-polluted evidence to derive accurate predictions. Existing works simulate GenAI-powered pollution at the claim level with stylistic rewriting to conceal linguistic cues, and ignore evidence-level pollution for such information-seeking applications. In this work, we investigate how polluted evidence affects the performance of existing OOC detectors, revealing a performance degradation of more than 9 percentage points. We propose two strategies, cross-modal evidence reranking and cross-modal claim-evidence reasoning, to address the challenges posed by polluted evidence. Extensive experiments on two benchmark datasets show that these strategies can effectively enhance the robustness of existing out-of-context detectors amidst polluted evidence.
2501.14729
HERMES: A Unified Self-Driving World Model for Simultaneous 3D Scene Understanding and Generation
cs.CV
Driving World Models (DWMs) have become essential for autonomous driving by enabling future scene prediction. However, existing DWMs are limited to scene generation and fail to incorporate scene understanding, which involves interpreting and reasoning about the driving environment. In this paper, we present a unified Driving World Model named HERMES. We seamlessly integrate 3D scene understanding and future scene evolution (generation) through a unified framework in driving scenarios. Specifically, HERMES leverages a Bird's-Eye View (BEV) representation to consolidate multi-view spatial information while preserving geometric relationships and interactions. We also introduce world queries, which incorporate world knowledge into BEV features via causal attention in the Large Language Model (LLM), enabling contextual enrichment for understanding and generation tasks. We conduct comprehensive studies on nuScenes and OmniDrive-nuScenes datasets to validate the effectiveness of our method. HERMES achieves state-of-the-art performance, reducing generation error by 32.4% and improving understanding metrics such as CIDEr by 8.0%. The model and code will be publicly released at https://github.com/LMD0311/HERMES.
2501.14731
From Critique to Clarity: A Pathway to Faithful and Personalized Code Explanations with Large Language Models
cs.SE cs.AI cs.CL
In the realm of software development, providing accurate and personalized code explanations is crucial for both technical professionals and business stakeholders. Technical professionals benefit from enhanced understanding and improved problem-solving skills, while business stakeholders gain insights into project alignments and transparency. Despite the potential, generating such explanations is often time-consuming and challenging. This paper presents an innovative approach that leverages the advanced capabilities of large language models (LLMs) to generate faithful and personalized code explanations. Our methodology integrates prompt enhancement, self-correction mechanisms, personalized content customization, and interaction with external tools, facilitated by collaboration among multiple LLM agents. We evaluate our approach using both automatic and human assessments, demonstrating that our method not only produces accurate explanations but also tailors them to individual user preferences. Our findings suggest that this approach significantly improves the quality and relevance of code explanations, offering a valuable tool for developers and stakeholders alike.
2501.14733
LLM as HPC Expert: Extending RAG Architecture for HPC Data
cs.DC cs.AI
High-Performance Computing (HPC) is crucial for performing advanced computational tasks, yet their complexity often challenges users, particularly those unfamiliar with HPC-specific commands and workflows. This paper introduces Hypothetical Command Embeddings (HyCE), a novel method that extends Retrieval-Augmented Generation (RAG) by integrating real-time, user-specific HPC data, enhancing accessibility to these systems. HyCE enriches large language models (LLM) with real-time, user-specific HPC information, addressing the limitations of fine-tuned models on such data. We evaluate HyCE using an automated RAG evaluation framework, where the LLM itself creates synthetic questions from the HPC data and serves as a judge, assessing the efficacy of the extended RAG with the evaluation metrics relevant for HPC tasks. Additionally, we tackle essential security concerns, including data privacy and command execution risks, associated with deploying LLMs in HPC environments. This solution provides a scalable and adaptable approach for HPC clusters to leverage LLMs as HPC expert, bridging the gap between users and the complex systems of HPC.
2501.14734
Research on the Application of Spark Streaming Real-Time Data Analysis System and large language model Intelligent Agents
cs.DC cs.AI
This study explores the integration of Agent AI with LangGraph to enhance real-time data analysis systems in big data environments. The proposed framework overcomes limitations of static workflows, inefficient stateful computations, and lack of human intervention by leveraging LangGraph's graph-based workflow construction and dynamic decision-making capabilities. LangGraph allows large language models (LLMs) to dynamically determine control flows, invoke tools, and assess the necessity of further actions, improving flexibility and efficiency. The system architecture incorporates Apache Spark Streaming, Kafka, and LangGraph to create a high-performance sentiment analysis system. LangGraph's capabilities include precise state management, dynamic workflow construction, and robust memory checkpointing, enabling seamless multi-turn interactions and context retention. Human-in-the-loop mechanisms are integrated to refine sentiment analysis, particularly in ambiguous or high-stakes scenarios, ensuring greater reliability and contextual relevance. Key features such as real-time state streaming, debugging via LangGraph Studio, and efficient handling of large-scale data streams make this framework ideal for adaptive decision-making. Experimental results confirm the system's ability to classify inquiries, detect sentiment trends, and escalate complex issues for manual review, demonstrating a synergistic blend of LLM capabilities and human oversight. This work presents a scalable, adaptable, and reliable solution for real-time sentiment analysis and decision-making, advancing the use of Agent AI and LangGraph in big data applications.
2501.14735
ARCEAK: An Automated Rule Checking Framework Enhanced with Architectural Knowledge
cs.SE cs.AI
Automated Rule Checking (ARC) plays a crucial role in advancing the construction industry by addressing the laborious, inconsistent, and error-prone nature of traditional model review conducted by industry professionals. Manual assessment against intricate sets of rules often leads to significant project delays and expenses. In response to these challenges, ARC offers a promising solution to improve efficiency and compliance in design within the construction sector. However, the main challenge of ARC lies in translating regulatory text into a format suitable for computer processing. Current methods for rule interpretation require extensive manual labor, thereby limiting their practicality. To address this issue, our study introduces a novel approach that decomposes ARC into two distinct tasks: rule information extraction and verification code generation. Leveraging generative pre-trained transformers, our method aims to streamline the interpretation of regulatory texts and simplify the process of generating model compliance checking code. Through empirical evaluation and case studies, we showcase the effectiveness and potential of our approach in automating code compliance checking, enhancing the efficiency and reliability of construction projects.
2501.14736
NEAT Algorithm-based Stock Trading Strategy with Multiple Technical Indicators Resonance
cs.NE cs.LG q-fin.PM
In this study, we applied the NEAT (NeuroEvolution of Augmenting Topologies) algorithm to stock trading using multiple technical indicators. Our approach focused on maximizing earning, avoiding risk, and outperforming the Buy & Hold strategy. We used progressive training data and a multi-objective fitness function to guide the evolution of the population towards these objectives. The results of our study showed that the NEAT model achieved similar returns to the Buy & Hold strategy, but with lower risk exposure and greater stability. We also identified some challenges in the training process, including the presence of a large number of unused nodes and connections in the model architecture. In future work, it may be worthwhile to explore ways to improve the NEAT algorithm and apply it to shorter interval data in order to assess the potential impact on performance.
2501.14737
EvalSVA: Multi-Agent Evaluators for Next-Gen Software Vulnerability Assessment
cs.SE cs.AI
Software Vulnerability (SV) assessment is a crucial process of determining different aspects of SVs (e.g., attack vectors and scope) for developers to effectively prioritize efforts in vulnerability mitigation. It presents a challenging and laborious process due to the complexity of SVs and the scarcity of labeled data. To mitigate the above challenges, we introduce EvalSVA, a multi-agent evaluators team to autonomously deliberate and evaluate various aspects of SV assessment. Specifically, we propose a multi-agent-based framework to simulate vulnerability assessment strategies in real-world scenarios, which employs multiple Large Language Models (LLMs) into an integrated group to enhance the effectiveness of SV assessment in the limited data. We also design diverse communication strategies to autonomously discuss and assess different aspects of SV. Furthermore, we construct a multi-lingual SV assessment dataset based on the new standard of CVSS, comprising 699, 888, and 1,310 vulnerability-related commits in C++, Python, and Java, respectively. Our experimental results demonstrate that EvalSVA averagely outperforms the 44.12\% accuracy and 43.29\% F1 for SV assessment compared with the previous methods. It shows that EvalSVA offers a human-like process and generates both reason and answer for SV assessment. EvalSVA can also aid human experts in SV assessment, which provides more explanation and details for SV assessment.
2501.14738
On strict ranking by pairwise comparisons
cs.IT math.IT
We attack the problem of getting a strict ranking (i.e. a ranking without equally ranked items) of $n$ items from a pairwise comparisons matrix. Basic structures are described, a first heuristical approach based on a condition, the $\mathcal{R}-$condition, is proposed. Analyzing the limits of this ranking procedure, we finish with a minimization problem which can be applied to a wider class of pairwise comparisons matrices. If solved, it produces consistent pairwise comparisons that produce a strict ranking.
2501.14739
Reproduction Research of FSA-Benchmark
cs.DC cs.LG
In the current landscape of big data, the reliability and performance of storage systems are essential to the success of various applications and services. as data volumes continue to grow exponentially, the complexity and scale of the storage infrastructures needed to manage this data also increase. a significant challenge faced by data centers and storage systems is the detection and management of fail-slow disks that experience a gradual decline in performance before ultimately failing. Unlike outright disk failures, fail-slow conditions can go undetected for prolonged periods, leading to considerable impacts on system performance and user experience.
2501.14741
On Design Choices in Similarity-Preserving Sparse Randomized Embeddings
cs.NE cs.LG q-bio.NC
Expand & Sparsify is a principle that is observed in anatomically similar neural circuits found in the mushroom body (insects) and the cerebellum (mammals). Sensory data are projected randomly to much higher-dimensionality (expand part) where only few the most strongly excited neurons are activated (sparsify part). This principle has been leveraged to design a FlyHash algorithm that forms similarity-preserving sparse embeddings, which have been found useful for such tasks as novelty detection, pattern recognition, and similarity search. Despite its simplicity, FlyHash has a number of design choices to be set such as preprocessing of the input data, choice of sparsifying activation function, and formation of the random projection matrix. In this paper, we explore the effect of these choices on the performance of similarity search with FlyHash embeddings. We find that the right combination of design choices can lead to drastic difference in the search performance.
2501.14742
Evaluating the effectiveness, reliability and efficiency of a multi-objective sequential optimization approach for building performance design
cs.NE math.OC
The complexity of performance-based building design stems from the evaluation of numerous candidate design options, driven by the plethora of variables, objectives, and constraints inherent in multi-disciplinary projects. This necessitates optimization approaches to support the identification of well performing designs while reducing the computational time of performance evaluation. In response, this paper proposes and evaluates a sequential approach for multi-objective design optimization of building geometry, fabric, HVAC system and controls for building performance. This approach involves sequential optimizations with optimal solutions from previous stages passed to the next. The performance of the sequential approach is benchmarked against a full factorial search, assessing its effectiveness in finding global optima, solution quality, reliability to scale and variations of problem formulations, and computational efficiency compared to the NSGA-II algorithm. 24 configurations of the sequential approach are tested on a multi-scale case study, simulating 874 to 4,147,200 design options for an office building, aiming to minimize energy demand while maintaining thermal comfort. A two-stage sequential process-(building geometry + fabric) and (HVAC system + controls) identified the same Pareto-optimal solutions as the full factorial search across all four scales and variations of problem formulations, demonstrating 100% effectiveness and reliability. This approach required 100,700 function evaluations, representing a 91.2% reduction in computational effort compared to the full factorial search. In contrast, NSGA-II achieved only 73.5% of the global optima with the same number of function evaluations. This research indicates that a sequential optimization approach is a highly efficient and robust alternative to the standard NSGA-II algorithm.
2501.14743
KVDirect: Distributed Disaggregated LLM Inference
cs.DC cs.LG cs.PF
Large Language Models (LLMs) have become the new foundation for many applications, reshaping human society like a storm. Disaggregated inference, which separates prefill and decode stages, is a promising approach to improving hardware utilization and service quality. However, due to inefficient inter-node communication, existing systems restrict disaggregated inference to a single node, limiting resource allocation flexibility and reducing service capacity. This paper introduces KVDirect, which optimizes KV cache transfer to enable a distributed disaggregated LLM inference. KVDirect achieves this through the following contributions. First, we propose a novel tensor-centric communication mechanism that reduces the synchronization overhead in traditional distributed GPU systems. Second, we design a custom communication library to support dynamic GPU resource scheduling and efficient KV cache transfer. Third, we introduce a pull-based KV cache transfer strategy that reduces GPU resource idling and improves latency. Finally, we implement KVDirect as an open-source LLM inference framework. Our evaluation demonstrates that KVDirect reduces per-request latency by 55% compared to the baseline across diverse workloads under the same resource constraints.
2501.14744
FSTA-SNN:Frequency-based Spatial-Temporal Attention Module for Spiking Neural Networks
cs.NE cs.CV cs.LG
Spiking Neural Networks (SNNs) are emerging as a promising alternative to Artificial Neural Networks (ANNs) due to their inherent energy efficiency. Owing to the inherent sparsity in spike generation within SNNs, the in-depth analysis and optimization of intermediate output spikes are often neglected. This oversight significantly restricts the inherent energy efficiency of SNNs and diminishes their advantages in spatiotemporal feature extraction, resulting in a lack of accuracy and unnecessary energy expenditure. In this work, we analyze the inherent spiking characteristics of SNNs from both temporal and spatial perspectives. In terms of spatial analysis, we find that shallow layers tend to focus on learning vertical variations, while deeper layers gradually learn horizontal variations of features. Regarding temporal analysis, we observe that there is not a significant difference in feature learning across different time steps. This suggests that increasing the time steps has limited effect on feature learning. Based on the insights derived from these analyses, we propose a Frequency-based Spatial-Temporal Attention (FSTA) module to enhance feature learning in SNNs. This module aims to improve the feature learning capabilities by suppressing redundant spike features.The experimental results indicate that the introduction of the FSTA module significantly reduces the spike firing rate of SNNs, demonstrating superior performance compared to state-of-the-art baselines across multiple datasets.
2501.14745
AI-Driven Health Monitoring of Distributed Computing Architecture: Insights from XGBoost and SHAP
cs.DC cs.LG
With the rapid development of artificial intelligence technology, its application in the optimization of complex computer systems is becoming more and more extensive. Edge computing is an efficient distributed computing architecture, and the health status of its nodes directly affects the performance and reliability of the entire system. In view of the lack of accuracy and interpretability of traditional methods in node health status judgment, this paper proposes a health status judgment method based on XGBoost and combines the SHAP method to analyze the interpretability of the model. Through experiments, it is verified that XGBoost has superior performance in processing complex features and nonlinear data of edge computing nodes, especially in capturing the impact of key features (such as response time and power consumption) on node status. SHAP value analysis further reveals the global and local importance of features, so that the model not only has high precision discrimination ability but also can provide intuitive explanations, providing data support for system optimization. Research shows that the combination of AI technology and computer system optimization can not only realize the intelligent monitoring of the health status of edge computing nodes but also provide a scientific basis for dynamic optimization scheduling, resource management and anomaly detection. In the future, with the in-depth development of AI technology, model dynamics, cross-node collaborative optimization and multimodal data fusion will become the focus of research, providing important support for the intelligent evolution of edge computing systems.
2501.14746
Neuromorphic Spiking Neural Network Based Classification of COVID-19 Spike Sequences
cs.NE cs.LG
The availability of SARS-CoV-2 (severe acute respiratory syndrome coronavirus 2) virus data post-COVID has reached exponentially to an enormous magnitude, opening research doors to analyze its behavior. Various studies are conducted by researchers to gain a deeper understanding of the virus, like genomic surveillance, etc, so that efficient prevention mechanisms can be developed. However, the unstable nature of the virus (rapid mutations, multiple hosts, etc) creates challenges in designing analytical systems for it. Therefore, we propose a neural network-based (NN) mechanism to perform an efficient analysis of the SARS-CoV-2 data, as NN portrays generalized behavior upon training. Moreover, rather than using the full-length genome of the virus, we apply our method to its spike region, as this region is known to have predominant mutations and is used to attach to the host cell membrane. In this paper, we introduce a pipeline that first converts the spike protein sequences into a fixed-length numerical representation and then uses Neuromorphic Spiking Neural Network to classify those sequences. We compare the performance of our method with various baselines using real-world SARS-CoV-2 spike sequence data and show that our method is able to achieve higher predictive accuracy compared to the recent baselines.
2501.14747
Enhancing Green Economy with Artificial Intelligence: Role of Energy Use and FDI in the United States
econ.GN cs.AI q-fin.EC
The escalating challenge of climate change necessitates an urgent exploration of factors influencing carbon emissions. This study contributes to the discourse by examining the interplay of technological, economic, and demographic factors on environmental sustainability. This study investigates the impact of artificial intelligence (AI) innovation, economic growth, foreign direct investment (FDI), energy consumption, and urbanization on CO2 emissions in the United States from 1990 to 2022. Employing the ARDL framework integrated with the STIRPAT model, the findings reveal a dual narrative: while AI innovation mitigates environmental stress, economic growth, energy use, FDI, and urbanization exacerbate environmental degradation. Unit root tests (ADF, PP, and DF-GLS) confirm mixed integration levels among variables, and the ARDL bounds test establishes long-term co-integration. The analysis highlights that AI innovation positively correlates with CO2 reduction when environmental safeguards are in place, whereas GDP growth, energy consumption, FDI, and urbanization intensify CO2 emissions. Robustness checks using FMOLS, DOLS, and CCR validate the ARDL findings. Additionally, Pairwise Granger causality tests reveal significant one-way causal links between CO2 emissions and economic growth, AI innovation, energy use, FDI, and urbanization. These relationships emphasize the critical role of AI-driven technological advancements, sustainable investments, and green energy in fostering ecological sustainability. The study suggests policy measures such as encouraging green FDI, advancing AI technologies, adopting sustainable energy practices, and implementing eco-friendly urban development to promote sustainable growth in the USA.
2501.14750
Engineering Carbon Credits Towards A Responsible FinTech Era: The Practices, Implications, and Future
cs.CY cs.LG
Carbon emissions significantly contribute to climate change, and carbon credits have emerged as a key tool for mitigating environmental damage and helping organizations manage their carbon footprint. Despite their growing importance across sectors, fully leveraging carbon credits remains challenging. This study explores engineering practices and fintech solutions to enhance carbon emission management. We first review the negative impacts of carbon emission non-disclosure, revealing its adverse effects on financial stability and market value. Organizations are encouraged to actively manage emissions and disclose relevant data to mitigate risks. Next, we analyze factors influencing carbon prices and review advanced prediction algorithms that optimize carbon credit purchasing strategies, reducing costs and improving efficiency. Additionally, we examine corporate carbon emission prediction models, which offer accurate performance assessments and aid in planning future carbon credit needs. By integrating carbon price and emission predictions, we propose research directions, including corporate carbon management cost forecasting. This study provides a foundation for future quantitative research on the financial and market impacts of carbon management practices and is the first systematic review focusing on computing solutions and engineering practices for carbon credits.
2501.14751
Optimizing LPB Algorithms using Simulated Annealing
cs.NE
Learner Performance-based Behavior using Simulated Annealing (LPBSA) is an improvement of the Learner Performance-based Behavior (LPB) algorithm. LPBSA, like LPB, has been proven to deal with single and complex problems. Simulated Annealing (SA) has been utilized as a powerful technique to optimize LPB. LPBSA has provided results that outperformed popular algorithms, like the Genetic Algorithm (GA), Particle Swarm Optimization (PSO), and even LPB. This study outlines the improved algorithm's working procedure by providing a main population and dividing it into Good and Bad populations and then applying crossover and mutation operators. When some individuals are born in the crossover stage, they have to go through the mutation process. Between these two steps, we have applied SA using the Metropolis Acceptance Criterion (MAC) to accept only the best and most useful individuals to be used in the next iteration. Finally, the outcomes demonstrate that the population is enhanced, leading to improved efficiency and validating the performance of LPBSA.
2501.14753
ABACUS: A FinOps Service for Cloud Cost Optimization
cs.DC cs.AI cs.NI cs.SE
In recent years, as more enterprises have moved their infrastructure to the cloud, significant challenges have emerged in achieving holistic cloud spend visibility and cost optimization. FinOps practices provide a way for enterprises to achieve these business goals by optimizing cloud costs and bringing accountability to cloud spend. This paper presents ABACUS - Automated Budget Analysis and Cloud Usage Surveillance, a FinOps solution for optimizing cloud costs by setting budgets, enforcing those budgets through blocking new deployments, and alerting appropriate teams if spending breaches a budget threshold. ABACUS also leverages best practices like Infrastructure-as-Code to alert engineering teams of the expected cost of deployment before resources are deployed in the cloud. Finally, future research directions are proposed to advance the state of the art in this important field.
2501.14755
Data-Juicer 2.0: Cloud-Scale Adaptive Data Processing for Foundation Models
cs.DC cs.AI
The burgeoning field of foundation models necessitates advanced data processing mechanisms capable of harnessing vast valuable data with varied types utilized by these models. Nevertheless, the current landscape presents unique challenges that traditional data processing frameworks cannot handle effectively, especially with multimodal intricacies. In response, we present Data-Juicer 2.0, a new system offering fruitful data processing capabilities backed by over a hundred operators spanning various modalities like text, image, audio, and video. With seamless compatibility and dedicated optimization to popular dataset hubs like Hugging Face and computing engines like Ray, Data-Juicer 2.0 enhances its predecessor in both usability, efficiency, and programmability. It features an easily accessible user interface layer that supports decoupled Python interactions, RESTful APIs, and conversational commands. Alongside this, it contains a core runtime layer optimized for adaptive execution and management across different dataset scales, processing demands, and computational environments, while shielding unnecessary system details. Extensive empirical evaluations demonstrate Data-Juicer 2.0's remarkable performance and scalability, highlighting its capability to efficiently process tens of billions of data samples with tens of thousands of CPU cores. The system is publicly available, actively maintained, and broadly adopted in diverse research endeavors, practical applications, and real-world products such as Alibaba Cloud PAI.
2501.14756
Towards An Automated AI Act FRIA Tool That Can Reuse GDPR's DPIA
cs.CY cs.AI
The AI Act introduces the obligation to conduct a Fundamental Rights Impact Assessment (FRIA), with the possibility to reuse a Data Protection Impact Assessment (DPIA), and requires the EU Commission to create of an automated tool to support the FRIA process. In this article, we provide our novel exploration of the DPIA and FRIA as information processes to enable the creation of automated tools. We first investigate the information involved in DPIA and FRIA, and then use this to align the two to state where a DPIA can be reused in a FRIA. We then present the FRIA as a 5-step process and discuss the role of an automated tool for each step. Our work provides the necessary foundation for creating and managing information for FRIA and supporting it through an automated tool as required by the AI Act.
2501.14759
LPBSA: Enhancing Optimization Efficiency through Learner Performance-based Behavior and Simulated Annealing
cs.NE math.OC
This study introduces the LPBSA, an advanced optimization algorithm that combines Learner Performance-based Behavior (LPB) and Simulated Annealing (SA) in a hybrid approach. Emphasizing metaheuristics, the LPBSA addresses and mitigates the challenges associated with traditional LPB methodologies, enhancing convergence, robustness, and adaptability in solving complex optimization problems. Through extensive evaluations using benchmark test functions, the LPBSA demonstrates superior performance compared to LPB and competes favorably with established algorithms such as PSO, FDO, LEO, and GA. Real-world applications underscore the algorithm's promise, with LPBSA outperforming the LEO algorithm in two tested scenarios. Based on the study results many test function results such as TF5 by recording (4.76762333) and some other test functions provided in the result section prove that LPBSA outperforms popular algorithms. This research highlights the efficacy of a hybrid approach in the ongoing evolution of optimization algorithms, showcasing the LPBSA's capacity to navigate diverse optimization landscapes and contribute significantly to addressing intricate optimization challenges.
2501.14762
Linked Data on Geo-annotated Events and Use Cases for the Resilience of Ukraine
cs.CY cs.SI
The mission of resilience of Ukrainian cities calls for international collaboration with the scientific community to increase the quality of information by identifying and integrating information from various news and social media sources. Linked Data technology can be used to unify, enrich, and integrate data from multiple sources. In our work, we focus on datasets about damaging events in Ukraine due to Russia's invasion between February 2022 and the end of April 2023. We convert two selected datasets to Linked Data and enrich them with additional geospatial information. Following that, we present an algorithm for the detection of identical events from different datasets. Our pipeline makes it easy to convert and enrich datasets to integrated Linked Data. The resulting dataset consists of 10K reported events covering damage to hospitals, schools, roads, residential buildings, etc. Finally, we demonstrate in use cases how our dataset can be applied to different scenarios for resilience purposes.
2501.14765
Hybrid Cooperative Co-Evolution Algorithm for Deadlock-prone Distributed Assembly Flowshop Scheduling with Limited buffers Using Petri nets
cs.DC cs.SY eess.SY
The distributed assembly flowshop scheduling problem (DAFSP) can be applied to immense manufacturing environments. In DAFSP, jobs are first processed in distributed flowshops, and then assembled into final products by an assembly machine, which usually has limited buffers in practical application. This limited capacity can lead to deadlocks, halting job completion and blocking the entire manufacturing process. However, existing scheduling methods fail to address these deadlocks in DAFSP effectively. As such, we develop a hybrid cooperative co-evolution (HCCE) algorithm for solving the deadlock-prone DAFSP by minimizing the makespan. For the first time, we use Petri nets to analyze the deadlocks in DAFSP and propose a Petri net-based deadlock amending method (IDAM), which is further integrated into HCCE to ensure the feasibility (i.e., deadlock-freeness) of solutions. Importantly, HCCE contains an elite archive (EAR) and two subpopulations. It uses the problem-specific operators for heuristic initialization and global-search. To enhance the quality and diversity of solutions, an information transfer mechanism (ITM) is developed among subpopulation and EAR, and four local-search operators are performed sequentially on each individual in EAR. Finally, comprehensive experiments demonstrate the effectiveness and superiority of the proposed HCCE algorithm.
2501.14766
Artificial Intelligence for Sustainable Urban Biodiversity: A Framework for Monitoring and Conservation
cs.CY cs.AI
The rapid expansion of urban areas challenges biodiversity conservation, requiring innovative ecosystem management. This study explores the role of Artificial Intelligence (AI) in urban biodiversity conservation, its applications, and a framework for implementation. Key findings show that: (a) AI enhances species detection and monitoring, achieving over 90% accuracy in urban wildlife tracking and invasive species management; (b) integrating data from remote sensing, acoustic monitoring, and citizen science enables large-scale ecosystem analysis; and (c) AI decision tools improve conservation planning and resource allocation, increasing prediction accuracy by up to 18.5% compared to traditional methods. The research presents an AI-Driven Framework for Urban Biodiversity Management, highlighting AI's impact on monitoring, conservation strategies, and ecological outcomes. Implementation strategies include: (a) standardizing data collection and model validation, (b) ensuring equitable AI access across urban contexts, and (c) developing ethical guidelines for biodiversity monitoring. The study concludes that integrating AI in urban biodiversity conservation requires balancing innovation with ecological wisdom and addressing data quality, socioeconomic disparities, and ethical concerns.
2501.14767
Leveraging Social Media Data and Artificial Intelligence for Improving Earthquake Response Efforts
cs.CY cs.AI cs.CL cs.IR cs.SI
The integration of social media and artificial intelligence (AI) into disaster management, particularly for earthquake response, represents a profound evolution in emergency management practices. In the digital age, real-time information sharing has reached unprecedented levels, with social media platforms emerging as crucial communication channels during crises. This shift has transformed traditional, centralized emergency services into more decentralized, participatory models of disaster situational awareness. Our study includes an experimental analysis of 8,900 social media interactions, including 2,920 posts and 5,980 replies on X (formerly Twitter), following a magnitude 5.1 earthquake in Oklahoma on February 2, 2024. The analysis covers data from the immediate aftermath and extends over the following seven days, illustrating the critical role of digital platforms in modern disaster response. The results demonstrate that social media platforms can be effectively used as real-time situational awareness tools, delivering critical information to society and authorities during emergencies.
2501.14768
Equation discovery framework EPDE: Towards a better equation discovery
cs.NE cs.AI cs.LG
Equation discovery methods hold promise for extracting knowledge from physics-related data. However, existing approaches often require substantial prior information that significantly reduces the amount of knowledge extracted. In this paper, we enhance the EPDE algorithm -- an evolutionary optimization-based discovery framework. In contrast to methods like SINDy, which rely on pre-defined libraries of terms and linearities, our approach generates terms using fundamental building blocks such as elementary functions and individual differentials. Within evolutionary optimization, we may improve the computation of the fitness function as is done in gradient methods and enhance the optimization algorithm itself. By incorporating multi-objective optimization, we effectively explore the search space, yielding more robust equation extraction, even when dealing with complex experimental data. We validate our algorithm's noise resilience and overall performance by comparing its results with those from the state-of-the-art equation discovery framework SINDy.
2501.14769
A survey on pioneering metaheuristic algorithms between 2019 and 2024
cs.NE cs.AI
This review examines over 150 new metaheuristics of the last six years (between 2019 and 2024), underscoring their profound influence and performance. Over the past three decades, more than 500 new metaheuristic algorithms have been proposed, with no slowdown in sight. An overwhelming abundance that complicates the process of selecting and assessing the most effective solutions for complex optimization challenges. Our evaluation centers on pivotal criteria, including annual citation metrics, the breadth of the addressed problem types, source code availability, user friendly parameter configurations, innovative mechanisms and operators, and approaches designed to mitigate traditional metaheuristic issues such as stagnation and premature convergence. We further explore recent high impact applications of the past six years' most influential 23 metahueristic algorithms, shedding light on their advantages and limitations, while identifying challenges and potential avenues for future research.
2501.14770
Optimizing SSD Caches for Cloud Block Storage Systems Using Machine Learning Approaches
cs.DC cs.LG cs.OS
The growing demand for efficient cloud storage solutions has led to the widespread adoption of Solid-State Drives (SSDs) for caching in cloud block storage systems. The management of data writes to SSD caches plays a crucial role in improving overall system performance, reducing latency, and extending the lifespan of storage devices. A critical challenge arises from the large volume of write-only data, which significantly impacts the performance of SSD caches when handled inefficiently. Specifically, writes that have not been read for a certain period may introduce unnecessary write traffic to the SSD cache without offering substantial benefits for cache performance. This paper proposes a novel approach to mitigate this issue by leveraging machine learning techniques to dynamically optimize the write policy in cloud-based storage systems. The proposed method identifies write-only data and selectively filters it out in real-time, thereby minimizing the number of unnecessary write operations and improving the overall performance of the cache system. Experimental results demonstrate that the proposed machine learning-based policy significantly outperforms traditional approaches by reducing the number of harmful writes and optimizing cache utilization. This solution is particularly suitable for cloud environments with varying and unpredictable workloads, where traditional cache management strategies often fall short.
2501.14771
Dynamic Adaptation in Data Storage: Real-Time Machine Learning for Enhanced Prefetching
cs.DC cs.LG cs.OS
The exponential growth of data storage demands has necessitated the evolution of hierarchical storage management strategies [1]. This study explores the application of streaming machine learning [3] to revolutionize data prefetching within multi-tiered storage systems. Unlike traditional batch-trained models, streaming machine learning [5] offers adaptability, real-time insights, and computational efficiency, responding dynamically to workload variations. This work designs and validates an innovative framework that integrates streaming classification models for predicting file access patterns, specifically the next file offset. Leveraging comprehensive feature engineering and real-time evaluation over extensive production traces, the proposed methodology achieves substantial improvements in prediction accuracy, memory efficiency, and system adaptability. The results underscore the potential of streaming models in real-time storage management, setting a precedent for advanced caching and tiering strategies.
2501.14772
DropMicroFluidAgents (DMFAs): Autonomous Droplet Microfluidic Research Framework Through Large Language Model Agents
cs.CY cs.AI
Applying Large language models (LLMs) within specific domains requires substantial adaptation to account for the unique terminologies, nuances, and context-specific challenges inherent to those areas. Here, we introduce DropMicroFluidAgents (DMFAs), an advanced language-driven framework leveraging state-of-the-art pre-trained LLMs. DMFAs employs LLM agents to perform two key functions: (1) delivering focused guidance, answers, and suggestions specific to droplet microfluidics and (2) generating machine learning models to optimise and automate the design of droplet microfluidic devices, including the creation of code-based computer-aided design (CAD) scripts to enable rapid and precise design execution. Experimental evaluations demonstrated that the integration of DMFAs with the LLAMA3.1 model yielded the highest accuracy of 76.15%, underscoring the significant performance enhancement provided by agent integration. This effect was particularly pronounced when DMFAs were paired with the GEMMA2 model, resulting in a 34.47% improvement in accuracy compared to the standalone GEMMA2 configuration. This study demonstrates the effective use of LLM agents in droplet microfluidics research as powerful tools for automating workflows, synthesising knowledge, optimising designs, and interacting with external systems. These capabilities enable their application across education and industrial support, driving greater efficiency in scientific discovery and innovation.
2501.14775
Hybrid Firefly-Genetic Algorithm for Single and Multi-dimensional 0-1 Knapsack Problems
cs.NE cs.AI
This paper addresses the challenges faced by algorithms, such as the Firefly Algorithm (FA) and the Genetic Algorithm (GA), in constrained optimization problems. While both algorithms perform well for unconstrained problems, their effectiveness diminishes when constraints are introduced due to limitations in exploration, exploitation, and constraint handling. To overcome these challenges, a hybrid FAGA algorithm is proposed, combining the strengths of both algorithms. The hybrid algorithm is validated by solving unconstrained benchmark functions and constrained optimization problems, including design engineering problems and combinatorial problems such as the 0-1 Knapsack Problem. The proposed algorithm delivers improved solution accuracy and computational efficiency compared to conventional optimization algorithm. This paper outlines the development and structure of the hybrid algorithm and demonstrates its effectiveness in handling complex optimization problems.
2501.14776
Green AI: Which Programming Language Consumes the Most?
cs.CY cs.AI cs.PL
AI is demanding an evergrowing portion of environmental resources. Despite their potential impact on AI environmental sustainability, the role that programming languages play in AI (in)efficiency is to date still unknown. With this study, we aim to understand the impact that programming languages can have on AI environmental sustainability. To achieve our goal, we conduct a controlled empirical experiment by considering five programming languages (C++, Java, Python, MATLAB, and R), seven AI algorithms (KNN, SVC, AdaBoost, decision tree, logistic regression, naive bayses, and random forest), three popular datasets, and the training and inference phases. The collected results show that programming languages have a considerable impact on AI environmental sustainability. Compiled and semi-compiled languages (C++, Java) consistently consume less than interpreted languages (Python, MATLAB, R), which require up to 54x more energy. Some languages are cumulatively more efficient in training, while others in inference. Which programming language consumes the most highly depends on the algorithm considered. Ultimately, algorithm implementation might be the most determining factor in Green AI, regardless of the language used. As conclusion, while making AI more environmentally sustainable is paramount, a trade-off between energy efficiency and implementation ease should always be considered. Green AI can be achieved without the need of completely disrupting the development practices and technologies currently in place.
2501.14777
Enhancing Supply Chain Resilience with Metaverse and ChatGPT Technologies
cs.CY cs.AI
Global supply lines have been severely disrupted by the COVID-19 epidemic and the conflict between Russia and Ukraine, which has sharply increased the price of commodities and generated inflation. These incidents highlight how critical it is to improve supply chain resilience (SCRES) in order to fend off unforeseen setbacks. Controlling both internal and external interruptions, such as transportation problems brought on by natural catastrophes and wars, is the responsibility of SCRES. Enhancing resilience in supply chains requires accurate and timely information transfer. Promising answers to these problems can be found in the Metaverse and ChatGPT, two new digital technologies. The Metaverse may imitate real-world situations and offer dynamic, real-time 3D representations of supply chain data by integrating blockchain, IoT, network connection, and computer power.Large-scale natural language processing model ChatGPT improves communication and data translation accuracy and speed. To manage risk and facilitate decision making in Supply Chain management, firms should increase information transmission, Speed and quality. This study aim to show the importance of ChatGPT and Metaverse technologies to improve SCRES, with an emphasis on the most important criteria for SCRES, and maturity factor that can influence directly the SC development.
2501.14778
Advancing Trustworthy AI for Sustainable Development: Recommendations for Standardising AI Incident Reporting
cs.CY cs.AI cs.HC
The increasing use of AI technologies has led to increasing AI incidents, posing risks and causing harm to individuals, organizations, and society. This study recognizes and addresses the lack of standardized protocols for reliably and comprehensively gathering such incident data crucial for preventing future incidents and developing mitigating strategies. Specifically, this study analyses existing open-access AI-incident databases through a systematic methodology and identifies nine gaps in current AI incident reporting practices. Further, it proposes nine actionable recommendations to enhance standardization efforts to address these gaps. Ensuring the trustworthiness of enabling technologies such as AI is necessary for sustainable digital transformation. Our research promotes the development of standards to prevent future AI incidents and promote trustworthy AI, thus facilitating achieving the UN sustainable development goals. Through international cooperation, stakeholders can unlock the transformative potential of AI, enabling a sustainable and inclusive future for all.
2501.14779
The Use of Generative Artificial Intelligence for Upper Secondary Mathematics Education Through the Lens of Technology Acceptance
cs.CY cs.AI cs.HC
This study investigated the students' perceptions of using Generative Artificial Intelligence (GenAI) in upper-secondary mathematics education. Data was collected from Finnish high school students to represent how key constructs of the Technology Acceptance Model (Perceived Usefulness, Perceived Ease of Use, Perceived Enjoyment, and Intention to Use) influence the adoption of AI tools. First, a structural equation model for a comparative study with a prior study was constructed and analyzed. Then, an extended model with the additional construct of Compatibility, which represents the alignment of AI tools with students' educational experiences and needs, was proposed and analyzed. The results demonstrated a strong influence of perceived usefulness on the intention to use GenAI, emphasizing the statistically significant role of perceived enjoyment in determining perceived usefulness and ease of use. The inclusion of compatibility improved the model's explanatory power, particularly in predicting perceived usefulness. This study contributes to a deeper understanding of how AI tools can be integrated into mathematics education and highlights key differences between the Finnish educational context and previous studies based on structural equation modeling.
2501.14780
Perspective Chapter: MOOCs in India: Evolution, Innovation, Impact, and Roadmap
cs.CY cs.AI cs.DL
With the largest population of the world and one of the highest enrolments in higher education, India needs efficient and effective means to educate its learners. India started focusing on open and digital education in 1980's and its efforts were escalated in 2009 through the NMEICT program of the Government of India. A study by the Government and FICCI in 2014 noted that India cannot meet its educational needs just by capacity building in brick and mortar institutions. It was decided that ongoing MOOCs projects under the umbrella of NMEICT will be further strengthened over its second (2017-21) and third (2021-26) phases. NMEICT now steers NPTEL or SWAYAM (India's MOOCs) and several digital learning projects including Virtual Labs, e-Yantra, Spoken Tutorial, FOSSEE, and National Digital Library on India - the largest digital education library in the world. Further, India embraced its new National Education Policy in 2020 to strongly foster online education. In this chapter, we take a deep look into the evolution of MOOCs in India, its innovations, its current status and impact, and the roadmap for the next decade to address its challenges and grow. AI-powered MOOCs is an emerging opportunity for India to lead MOOCs worldwide.
2501.14784
DeServe: Towards Affordable Offline LLM Inference via Decentralization
cs.DC cs.AI
The rapid growth of generative AI and its integration into everyday workflows have significantly increased the demand for large language model (LLM) inference services. While proprietary models remain popular, recent advancements in open-source LLMs have positioned them as strong contenders. However, deploying these models is often constrained by the high costs and limited availability of GPU resources. In response, this paper presents the design of a decentralized offline serving system for LLM inference. Utilizing idle GPU resources, our proposed system, DeServe, decentralizes access to LLMs at a lower cost. DeServe specifically addresses key challenges in optimizing serving throughput in high-latency network environments. Experiments demonstrate that DeServe achieves a 6.7x-12.6x improvement in throughput over existing serving system baselines in such conditions.
2501.14785
ED-Filter: Dynamic Feature Filtering for Eating Disorder Classification
stat.ML cs.AI cs.LG cs.SI
Eating disorders (ED) are critical psychiatric problems that have alarmed the mental health community. Mental health professionals are increasingly recognizing the utility of data derived from social media platforms such as Twitter. However, high dimensionality and extensive feature sets of Twitter data present remarkable challenges for ED classification. To overcome these hurdles, we introduce a novel method, an informed branch and bound search technique known as ED-Filter. This strategy significantly improves the drawbacks of conventional feature selection algorithms such as filters and wrappers. ED-Filter iteratively identifies an optimal set of promising features that maximize the eating disorder classification accuracy. In order to adapt to the dynamic nature of Twitter ED data, we enhance the ED-Filter with a hybrid greedy-based deep learning algorithm. This algorithm swiftly identifies sub-optimal features to accommodate the ever-evolving data landscape. Experimental results on Twitter eating disorder data affirm the effectiveness and efficiency of ED-Filter. The method demonstrates significant improvements in classification accuracy and proves its value in eating disorder detection on social media platforms.
2501.14786
Punch Out Model Synthesis: A Stochastic Algorithm for Constraint Based Tiling Generation
cs.DC cs.LG
As an artistic aid in tiled level design, Constraint Based Tiling Generation (CBTG) algorithms can help to automatically create level realizations from a set of tiles and placement constraints. Merrell's Modify in Blocks Model Synthesis (MMS) and Gumin's Wave Function Collapse (WFC) have been proposed as Constraint Based Tiling Generation (CBTG) algorithms that work well for many scenarios but have limitations in problem size, problem setup and solution biasing. We present Punch Out Model Synthesis (POMS), a Constraint Based Tiling Generation algorithm, that can handle large problem sizes, requires minimal assumptions for setup and can help mitigate solution biasing. POMS attempts to resolve indeterminate grid regions by trying to progressively realize sub-blocks, performing a stochastic boundary erosion on previously resolved regions should sub-block resolution fail. We highlight the results of running a reference implementation on different tile sets and discuss a tile correlation length, implied by the tile constraints, and its role in choosing an appropriate block size to aid POMS in successfully finding grid realizations.
2501.14787
Matrix Calculus (for Machine Learning and Beyond)
math.HO cs.LG cs.NA math.NA stat.ML
This course, intended for undergraduates familiar with elementary calculus and linear algebra, introduces the extension of differential calculus to functions on more general vector spaces, such as functions that take as input a matrix and return a matrix inverse or factorization, derivatives of ODE solutions, and even stochastic derivatives of random functions. It emphasizes practical computational applications, such as large-scale optimization and machine learning, where derivatives must be re-imagined in order to be propagated through complicated calculations. The class also discusses efficiency concerns leading to "adjoint" or "reverse-mode" differentiation (a.k.a. "backpropagation"), and gives a gentle introduction to modern automatic differentiation (AD) techniques.
2501.14788
Methods to Increase the Amount of Data for Speech Recognition for Low Resource Languages
cs.SD cs.CL eess.AS
This study explores methods to increase data volume for low-resource languages using techniques such as crowdsourcing, pseudo-labeling, advanced data preprocessing and various permissive data sources such as audiobooks, Common Voice, YouTube. While these methods are well-explored for highresource languages, their application for low-resource languages remains underexplored. Using Armenian and Georgian as case studies, we demonstrate how linguistic and resource-specific characteristics influence the success of these methods. This work provides practical guidance for researchers to choose cost-effective and quality-driven dataset extension strategies for low-resource languages. The key takeaway from various data extension approaches is that paid crowd-sourcing offers the best balance between cost and quality, outperforming volunteer crowd-sourcing, open-source audiobooks, and unlabeled data usage. Ablation study shows that models trained on the expanded datasets outperform existing baselines and achieve 5.73% for Gergian and 9.9% for Armenian ASR word error rate using a relatively small FastConformer architecture. We open-sourced both the Armenian and Georgian models to allow further research and practical applications.
2501.14790
Towards Dynamic Neural Communication and Speech Neuroprosthesis Based on Viseme Decoding
q-bio.NC cs.AI cs.SD eess.AS
Decoding text, speech, or images from human neural signals holds promising potential both as neuroprosthesis for patients and as innovative communication tools for general users. Although neural signals contain various information on speech intentions, movements, and phonetic details, generating informative outputs from them remains challenging, with mostly focusing on decoding short intentions or producing fragmented outputs. In this study, we developed a diffusion model-based framework to decode visual speech intentions from speech-related non-invasive brain signals, to facilitate face-to-face neural communication. We designed an experiment to consolidate various phonemes to train visemes of each phoneme, aiming to learn the representation of corresponding lip formations from neural signals. By decoding visemes from both isolated trials and continuous sentences, we successfully reconstructed coherent lip movements, effectively bridging the gap between brain signals and dynamic visual interfaces. The results highlight the potential of viseme decoding and talking face reconstruction from human neural signals, marking a significant step toward dynamic neural communication systems and speech neuroprosthesis for patients.
2501.14794
HeteroLLM: Accelerating Large Language Model Inference on Mobile SoCs platform with Heterogeneous AI Accelerators
cs.DC cs.AI cs.LG
With the rapid advancement of artificial intelligence technologies such as ChatGPT, AI agents and video generation,contemporary mobile systems have begun integrating these AI capabilities on local devices to enhance privacy and reduce response latency. To meet the computational demands of AI tasks, current mobile SoCs are equipped with diverse AI accelerators, including GPUs and Neural Processing Units (NPUs). However, there has not been a comprehensive characterization of these heterogeneous processors, and existing designs typically only leverage a single AI accelerator for LLM inference, leading to suboptimal use of computational resources and memory bandwidth. In this paper, we first summarize key performance characteristics of mobile SoC, including heterogeneous processors, unified memory, synchronization, etc. Drawing on these observations, we propose different tensor partition strategies to fulfill the distinct requirements of the prefill and decoding phases. We further design a fast synchronization mechanism that leverages the unified memory address provided by mobile SoCs. By employing these techniques, we present HeteroLLM, the fastest LLM inference engine in mobile devices which supports both layer-level and tensor-level heterogeneous execution. Evaluation results show that HeteroLLM achieves 9.99 and 4.36 performance improvement over other mobile-side LLM inference engines: MLC and MNN.
2501.14802
DNN-Powered MLOps Pipeline Optimization for Large Language Models: A Framework for Automated Deployment and Resource Management
cs.DC cs.LG
The exponential growth in the size and complexity of Large Language Models (LLMs) has introduced unprecedented challenges in their deployment and operational management. Traditional MLOps approaches often fail to efficiently handle the scale, resource requirements, and dynamic nature of these models. This research presents a novel framework that leverages Deep Neural Networks (DNNs) to optimize MLOps pipelines specifically for LLMs. Our approach introduces an intelligent system that automates deployment decisions, resource allocation, and pipeline optimization while maintaining optimal performance and cost efficiency. Through extensive experimentation across multiple cloud environments and deployment scenarios, we demonstrate significant improvements: 40% enhancement in resource utilization, 35% reduction in deployment latency, and 30% decrease in operational costs compared to traditional MLOps approaches. The framework's ability to adapt to varying workloads and automatically optimize deployment strategies represents a significant advancement in automated MLOps management for large-scale language models. Our framework introduces several novel components including a multi-stream neural architecture for processing heterogeneous operational metrics, an adaptive resource allocation system that continuously learns from deployment patterns, and a sophisticated deployment orchestration mechanism that automatically selects optimal strategies based on model characteristics and environmental conditions. The system demonstrates robust performance across various deployment scenarios, including multi-cloud environments, high-throughput production systems, and cost-sensitive deployments. Through rigorous evaluation using production workloads from multiple organizations, we validate our approach's effectiveness in reducing operational complexity while improving system reliability and cost efficiency.
2501.14808
HyGen: Efficient LLM Serving via Elastic Online-Offline Request Co-location
cs.DC cs.LG
Large language models (LLMs) have facilitated a wide range of applications with distinct service-level objectives (SLOs), from latency-sensitive online tasks like interactive chatbots to throughput-oriented offline workloads like document summarization. The existing deployment model, which dedicates machines to each workload, simplifies SLO management but often leads to poor resource utilization. This paper introduces HyGen, an interference-aware LLM serving system that enables efficient co-location of online and offline workloads while preserving latency requirements. HyGen incorporates two key innovations: (1) performance control mechanisms, including a latency predictor to estimate batch execution time and an SLO-aware profiler to quantify latency interference, and (2) SLO-aware offline scheduling policies that maximize serving throughput and prevent starvation, without compromising online serving latency. Our evaluation on production workloads shows that HyGen achieves up to 3.87x overall throughput and 5.84x offline throughput gains over online and hybrid serving baselines, respectively, while strictly satisfying latency SLOs.
2501.14809
Towards Foundation Models: Evaluation of Geoscience Artificial Intelligence with Uncertainty
cs.LG cs.AI physics.geo-ph
Artificial intelligence (AI) has transformed the geoscience community with deep learning models (DLMs) that are trained to complete specific tasks within workflows. This success has led to the development of geoscience foundation models (FMs), which promise to accomplish multiple tasks within a workflow or replace the workflow altogether. However, lack of robust evaluation frameworks, even for traditional DLMs, leaves the geoscience community ill prepared for the inevitable adoption of FMs. We address this gap by designing an evaluation framework that jointly incorporates three crucial aspects to current DLMs and future FMs: performance uncertainty, learning efficiency, and overlapping training-test data splits. To target the three aspects, we meticulously construct the training, validation, and test splits using clustering methods tailored to geoscience data and enact an expansive training design to segregate performance uncertainty arising from stochastic training processes and random data sampling. The framework's ability to guard against misleading declarations of model superiority is demonstrated through evaluation of PhaseNet, a popular seismic phase picking DLM, under 3 training approaches. Furthermore, we show how the performance gains due to overlapping training-test data can lead to biased FM evaluation. Our framework helps practitioners choose the best model for their problem and set performance expectations by explicitly analyzing model performance at varying budgets of training data.
2501.14813
Dissertation Machine Learning in Materials Science -- A case study in Carbon Nanotube field effect transistors
physics.app-ph cond-mat.mes-hall cs.LG physics.data-an
In this thesis, I explored the use of several machine learning techniques, including neural networks, simulation-based inference, and generative flow networks, on predicting CNTFETs performance, probing the conductivity properties of CNT network, and generating CNTFETs processing information for target performance.
2501.14815
A VM-HDL Co-Simulation Framework for Systems with PCIe-Connected FPGAs
cs.DC cs.AI cs.AR cs.NI
PCIe-connected FPGAs are gaining popularity as an accelerator technology in data centers. However, it is challenging to jointly develop and debug host software and FPGA hardware. Changes to the hardware design require a time-consuming FPGA synthesis process, and modification to the software, especially the operating system and device drivers, can frequently cause the system to hang, without providing enough information for debugging. The combination of these problems results in long debug iterations and a slow development process. To overcome these problems, we designed a VM-HDL co-simulation framework, which is capable of running the same software, operating system, and hardware designs as the target physical system, while providing full visibility and significantly shorter debug iterations.
2501.14816
Jump Point Search Pathfinding in 4-connected Grids
cs.RO
This work introduces JPS4, a novel pathfinding algorithm for 4-connected grid maps. JPS4 builds upon the Jump Point Search (JPS8) algorithm, originally designed for 8-connected environments. To achieve efficient pathfinding on 4-connected grids, JPS4 employs a canonical ordering and a successor function that enable online graph pruning. This reduces the search space by minimizing unnecessary node expansions. The core concept of JPS4 as well as JPS8 lies in the utilization of jump points. Strategically placed at obstacle corners, jump points prevent the search from overlooking crucial sections of the state space. They essentially reinitialize the canonical ordering, allowing exploration beyond obstacles. This mechanism ensures JPS4 finds optimal paths even in complex environments. The paper further explores the optimality of JPS4 and compares its performance against the established A* algorithm on various grid maps. Benchmarking results demonstrate that JPS4 significantly outperforms A* in scenarios with high obstacle density. However, A* remains more efficient on open maps. Overall, JPS4 presents itself as a promising alternative to A* for pathfinding on 4-connected grids, particularly applicable in video game development.
2501.14817
A Cutting Mechanics-based Machine Learning Modeling Method to Discover Governing Equations of Machining Dynamics
cs.LG cs.CE
This paper proposes a cutting mechanics-based machine learning (CMML) modeling method to discover governing equations of machining dynamics. The main idea of CMML design is to integrate existing physics in cutting mechanics and unknown physics in data to achieve automated model discovery, with the potential to advance machining modeling. Based on existing physics in cutting mechanics, CMML first establishes a general modeling structure governing machining dynamics, that is represented by a set of unknown differential algebraic equations. CMML can therefore achieve data-driven discovery of these unknown equations through effective cutting mechanics-based nonlinear learning function space design and discrete optimization-based learning algorithm. Experimentally verified time domain simulation of milling is used to validate the proposed modeling method. Numerical results show CMML can discover the exact milling dynamics models with process damping and edge force from noisy data. This indicates that CMML has the potential to be used for advancing machining modeling in practice with the development of effective metrology systems.
2501.14818
Eagle 2: Building Post-Training Data Strategies from Scratch for Frontier Vision-Language Models
cs.CV cs.AI cs.LG
Recently, promising progress has been made by open-source vision-language models (VLMs) in bringing their capabilities closer to those of proprietary frontier models. However, most open-source models only publish their final model weights, leaving the critical details of data strategies and implementation largely opaque. In this work, we address VLM post-training from a data-centric perspective, showing the key role of data strategy in developing frontier VLMs. By studying and building our post-training data strategy from scratch, we share detailed insights into the development processes, aiming to benefit the development of competitive models for the open-source community. Our introduced data strategy, together with training recipes and model design, leads to a family of performant VLMs named Eagle2. Specifically, Eagle2-9B achieves state-of-the-art results across various multimodal benchmarks, matching certain competitive models with up to 70B parameters.
2501.14819
A Comprehensive Mathematical and System-Level Analysis of Autonomous Vehicle Timelines
cs.MA cs.RO
Fully autonomous vehicles (AVs) continue to spark immense global interest, yet predictions on when they will operate safely and broadly remain heavily debated. This paper synthesizes two distinct research traditions: computational complexity and algorithmic constraints versus reliability growth modeling and real-world testing to form an integrated, quantitative timeline for future AV deployment. We propose a mathematical framework that unifies NP-hard multi-agent path planning analyses, high-performance computing (HPC) projections, and extensive Crow-AMSAA reliability growth calculations, factoring in operational design domain (ODD) variations, severity, and partial vs. full domain restrictions. Through category-specific case studies (e.g., consumer automotive, robo-taxis, highway trucking, industrial and defense applications), we show how combining HPC limitations, safety demonstration requirements, production/regulatory hurdles, and parallel/serial test strategies can push out the horizon for universal Level 5 deployment by up to several decades. Conversely, more constrained ODDs; like fenced industrial sites or specialized defense operations; may see autonomy reach commercial viability in the near-to-medium term. Our findings illustrate that while targeted domains can achieve automated service sooner, widespread driverless vehicles handling every environment remain far from realized. This paper thus offers a unique and rigorous perspective on why AV timelines extend well beyond short-term optimism, underscoring how each dimension of complexity and reliability imposes its own multi-year delays. By quantifying these constraints and exploring potential accelerators (e.g., advanced AI hardware, infrastructure up-grades), we provide a structured baseline for researchers, policymakers, and industry stakeholders to more accurately map their expectations and investments in AV technology.
2501.14822
Controlling Ensemble Variance in Diffusion Models: An Application for Reanalyses Downscaling
stat.AP cs.AI cs.LG
In recent years, diffusion models have emerged as powerful tools for generating ensemble members in meteorology. In this work, we demonstrate that a Denoising Diffusion Implicit Model (DDIM) can effectively control ensemble variance by varying the number of diffusion steps. Introducing a theoretical framework, we relate diffusion steps to the variance expressed by the reverse diffusion process. Focusing on reanalysis downscaling, we propose an ensemble diffusion model for the full ERA5-to-CERRA domain, generating variance-calibrated ensemble members for wind speed at full spatial and temporal resolution. Our method aligns global mean variance with a reference ensemble dataset and ensures spatial variance is distributed in accordance with observed meteorological variability. Additionally, we address the lack of ensemble information in the CARRA dataset, showcasing the utility of our approach for efficient, high-resolution ensemble generation.
2501.14823
Quantifying Energy and Cost Benefits of Hybrid Edge Cloud: Analysis of Traditional and Agentic Workloads
cs.DC cs.AI
This paper examines the workload distribution challenges in centralized cloud systems and demonstrates how Hybrid Edge Cloud (HEC) [1] mitigates these inefficiencies. Workloads in cloud environments often follow a Pareto distribution, where a small percentage of tasks consume most resources, leading to bottlenecks and energy inefficiencies. By analyzing both traditional workloads reflective of typical IoT and smart device usage and agentic workloads, such as those generated by AI agents, robotics, and autonomous systems, this study quantifies the energy and cost savings enabled by HEC. Our findings reveal that HEC achieves energy savings of up to 75% and cost reductions exceeding 80%, even in resource-intensive agentic scenarios. These results highlight the critical role of HEC in enabling scalable, cost-effective, and sustainable computing for the next generation of intelligent systems.
2501.14824
A causal learning approach to in-orbit inertial parameter estimation for multi-payload deployers
eess.SY astro-ph.IM cs.LG cs.RO cs.SY
This paper discusses an approach to inertial parameter estimation for the case of cargo carrying spacecraft that is based on causal learning, i.e. learning from the responses of the spacecraft, under actuation. Different spacecraft configurations (inertial parameter sets) are simulated under different actuation profiles, in order to produce an optimised time-series clustering classifier that can be used to distinguish between them. The actuation is comprised of finite sequences of constant inputs that are applied in order, based on typical actuators available. By learning from the system's responses across multiple input sequences, and then applying measures of time-series similarity and F1-score, an optimal actuation sequence can be chosen either for one specific system configuration or for the overall set of possible configurations. This allows for both estimation of the inertial parameter set without any prior knowledge of state, as well as validation of transitions between different configurations after a deployment event. The optimisation of the actuation sequence is handled by a reinforcement learning model that uses the proximal policy optimisation (PPO) algorithm, by repeatedly trying different sequences and evaluating the impact on classifier performance according to a multi-objective metric.
2501.14826
Multi-Modality Transformer for E-Commerce: Inferring User Purchase Intention to Bridge the Query-Product Gap
cs.IR cs.AI cs.LG
E-commerce click-stream data and product catalogs offer critical user behavior insights and product knowledge. This paper propose a multi-modal transformer termed as PINCER, that leverages the above data sources to transform initial user queries into pseudo-product representations. By tapping into these external data sources, our model can infer users' potential purchase intent from their limited queries and capture query relevant product features. We demonstrate our model's superior performance over state-of-the-art alternatives on e-commerce online retrieval in both controlled and real-world experiments. Our ablation studies confirm that the proposed transformer architecture and integrated learning strategies enable the mining of key data sources to infer purchase intent, extract product features, and enhance the transformation pipeline from queries to more accurate pseudo-product representations.
2501.14828
An Ensemble Model with Attention Based Mechanism for Image Captioning
cs.CV cs.AI
Image captioning creates informative text from an input image by creating a relationship between the words and the actual content of an image. Recently, deep learning models that utilize transformers have been the most successful in automatically generating image captions. The capabilities of transformer networks have led to notable progress in several activities related to vision. In this paper, we thoroughly examine transformer models, emphasizing the critical role that attention mechanisms play. The proposed model uses a transformer encoder-decoder architecture to create textual captions and a deep learning convolutional neural network to extract features from the images. To create the captions, we present a novel ensemble learning framework that improves the richness of the generated captions by utilizing several deep neural network architectures based on a voting mechanism that chooses the caption with the highest bilingual evaluation understudy (BLEU) score. The proposed model was evaluated using publicly available datasets. Using the Flickr8K dataset, the proposed model achieved the highest BLEU-[1-3] scores with rates of 0.728, 0.495, and 0.323, respectively. The suggested model outperformed the latest methods in Flickr30k datasets, determined by BLEU-[1-4] scores with rates of 0.798, 0.561, 0.387, and 0.269, respectively. The model efficacy was also obtained by the Semantic propositional image caption evaluation (SPICE) metric with a scoring rate of 0.164 for the Flicker8k dataset and 0.387 for the Flicker30k. Finally, ensemble learning significantly advances the process of image captioning and, hence, can be leveraged in various applications across different domains.
2501.14830
Sharp exact recovery threshold for two-community Euclidean random graphs
cs.SI math.PR
This paper considers the problem of label recovery in random graphs and matrices. Motivated by transitive behavior in real-world networks (i.e., ``the friend of my friend is my friend''), a recent line of work considers spatially-embedded networks, which exhibit transitive behavior. In particular, the Geometric Hidden Community Model (GHCM), introduced by Gaudio, Guan, Niu, and Wei, models a network as a labeled Poisson point process where every pair of vertices is associated with a pairwise observation whose distribution depends on the labels and positions of the vertices. The GHCM is in turn a generalization of the Geometric SBM (proposed by Baccelli and Sankararaman). Gaudio et al. provided a threshold below which exact recovery is information-theoretically impossible. Above the threshold, they provided a linear-time algorithm that succeeds in exact recovery under a certain ``distinctness-of-distributions'' assumption, which they conjectured to be unnecessary. In this paper, we partially resolve the conjecture by showing that the threshold is indeed tight for the two-community GHCM. We provide a two-phase, linear-time algorithm that explores the spatial graph in a data-driven manner in Phase I to yield an almost exact labeling, which is refined to achieve exact recovery in Phase II. Our results extend achievability to geometric formulations of well-known inference problems, such as the planted dense subgraph problem and submatrix localization, in which the distinctness-of-distributions assumption does not hold.
2501.14836
Symbolic Knowledge Extraction and Injection with Sub-symbolic Predictors: A Systematic Literature Review
cs.AI cs.LG cs.LO
In this paper we focus on the opacity issue of sub-symbolic machine learning predictors by promoting two complementary activities, namely, symbolic knowledge extraction (SKE) and injection (SKI) from and into sub-symbolic predictors. We consider as symbolic any language being intelligible and interpretable for both humans and computers. Accordingly, we propose general meta-models for both SKE and SKI, along with two taxonomies for the classification of SKE and SKI methods. By adopting an explainable artificial intelligence (XAI) perspective, we highlight how such methods can be exploited to mitigate the aforementioned opacity issue. Our taxonomies are attained by surveying and classifying existing methods from the literature, following a systematic approach, and by generalising the results of previous surveys targeting specific sub-topics of either SKE or SKI alone. More precisely, we analyse 132 methods for SKE and 117 methods for SKI, and we categorise them according to their purpose, operation, expected input/output data and predictor types. For each method, we also indicate the presence/lack of runnable software implementations. Our work may be of interest for data scientists aiming at selecting the most adequate SKE/SKI method for their needs, and also work as suggestions for researchers interested in filling the gaps of the current state of the art, as well as for developers willing to implement SKE/SKI-based technologies.
2501.14837
A Semiparametric Bayesian Method for Instrumental Variable Analysis with Partly Interval-Censored Time-to-Event Outcome
stat.ME cs.LG stat.AP stat.CO stat.ML
This paper develops a semiparametric Bayesian instrumental variable analysis method for estimating the causal effect of an endogenous variable when dealing with unobserved confounders and measurement errors with partly interval-censored time-to-event data, where event times are observed exactly for some subjects but left-censored, right-censored, or interval-censored for others. Our method is based on a two-stage Dirichlet process mixture instrumental variable (DPMIV) model which simultaneously models the first-stage random error term for the exposure variable and the second-stage random error term for the time-to-event outcome using a bivariate Gaussian mixture of the Dirichlet process (DPM) model. The DPM model can be broadly understood as a mixture model with an unspecified number of Gaussian components, which relaxes the normal error assumptions and allows the number of mixture components to be determined by the data. We develop an MCMC algorithm for the DPMIV model tailored for partly interval-censored data and conduct extensive simulations to assess the performance of our DPMIV method in comparison with some competing methods. Our simulations revealed that our proposed method is robust under different error distributions and can have superior performance over its parametric counterpart under various scenarios. We further demonstrate the effectiveness of our approach on an UK Biobank data to investigate the causal effect of systolic blood pressure on time-to-development of cardiovascular disease from the onset of diabetes mellitus.
2501.14844
Unmasking Conversational Bias in AI Multiagent Systems
cs.CL cs.AI cs.MA
Detecting biases in the outputs produced by generative models is essential to reduce the potential risks associated with their application in critical settings. However, the majority of existing methodologies for identifying biases in generated text consider the models in isolation and neglect their contextual applications. Specifically, the biases that may arise in multi-agent systems involving generative models remain under-researched. To address this gap, we present a framework designed to quantify biases within multi-agent systems of conversational Large Language Models (LLMs). Our approach involves simulating small echo chambers, where pairs of LLMs, initialized with aligned perspectives on a polarizing topic, engage in discussions. Contrary to expectations, we observe significant shifts in the stance expressed in the generated messages, particularly within echo chambers where all agents initially express conservative viewpoints, in line with the well-documented political bias of many LLMs toward liberal positions. Crucially, the bias observed in the echo-chamber experiment remains undetected by current state-of-the-art bias detection methods that rely on questionnaires. This highlights a critical need for the development of a more sophisticated toolkit for bias detection and mitigation for AI multi-agent systems. The code to perform the experiments is publicly available at https://anonymous.4open.science/r/LLMsConversationalBias-7725.
2501.14846
Wormhole Memory: A Rubik's Cube for Cross-Dialogue Retrieval
cs.LG cs.AI cs.CL
In view of the gap in the current large language model in sharing memory across dialogues, this research proposes a wormhole memory module (WMM) to realize memory as a Rubik's cube that can be arbitrarily retrieved between different dialogues. Through simulation experiments, the researcher built an experimental framework based on the Python environment and used setting memory barriers to simulate the current situation where memories between LLMs dialogues are difficult to share. The CoQA development data set was imported into the experiment, and the feasibility of its cross-dialogue memory retrieval function was verified for WMM's nonlinear indexing and dynamic retrieval, and a comparative analysis was conducted with the capabilities of Titans and MemGPT memory modules. Experimental results show that WMM demonstrated the ability to retrieve memory across dialogues and the stability of quantitative indicators in eight experiments. It contributes new technical approaches to the optimization of memory management of LLMs and provides experience for the practical application in the future.