id
stringlengths
9
16
title
stringlengths
4
278
categories
stringlengths
5
104
abstract
stringlengths
6
4.09k
2501.05729
ExPO: Explainable Phonetic Trait-Oriented Network for Speaker Verification
cs.SD cs.AI eess.AS
In speaker verification, we use computational method to verify if an utterance matches the identity of an enrolled speaker. This task is similar to the manual task of forensic voice comparison, where linguistic analysis is combined with auditory measurements to compare and evaluate voice samples. Despite much success, we have yet to develop a speaker verification system that offers explainable results comparable to those from manual forensic voice comparison. A novel approach, Explainable Phonetic Trait-Oriented (ExPO) network, is proposed in this paper to introduce the speaker's phonetic trait which describes the speaker's characteristics at the phonetic level, resembling what forensic comparison does. ExPO not only generates utterance-level speaker embeddings but also allows for fine-grained analysis and visualization of phonetic traits, offering an explainable speaker verification process. Furthermore, we investigate phonetic traits from within-speaker and between-speaker variation perspectives to determine which trait is most effective for speaker verification, marking an important step towards explainable speaker verification. Our code is available at https://github.com/mmmmayi/ExPO.
2501.05730
Element-wise Attention Is All You Need
cs.LG cs.AI
The self-attention (SA) mechanism has demonstrated superior performance across various domains, yet it suffers from substantial complexity during both training and inference. The next-generation architecture, aiming at retaining the competitive performance of SA while achieving low-cost inference and efficient long-sequence training, primarily focuses on three approaches: linear attention, linear RNNs, and state space models. Although these approaches achieve reduced complexity than SA, they all have built-in performance degradation factors, such as diminished “spikiness” and compression of historical information. In contrast to these approaches, we propose a novel element-wise attention mechanism, which uses the element-wise squared Euclidean distance, instead of the dot product operation, to compute similarity and approximates the quadratic complexity term $\exp(q_{ic}k_{jc})$ with a Taylor polynomial. This design achieves remarkable efficiency: during training, the element-wise attention has a complexity of $\mathcal{O}(tLD)$, making long-sequence training both computationally and memory efficient, where $L$ is the sequence length, $D$ is the feature dimension, and $t$ is the highest order of the polynomial; during inference, it can be reformulated as recurrent neural networks, achieving a inference complexity of $\mathcal{O}(tD)$. Furthermore, the element-wise attention circumvents the performance degradation factors present in these approaches and achieves performance comparable to SA in both causal and non-causal forms.
2501.05731
Diving Deep: Forecasting Sea Surface Temperatures and Anomalies
cs.LG physics.ao-ph stat.AP
This overview paper details the findings from the Diving Deep: Forecasting Sea Surface Temperatures and Anomalies Challenge at the European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases (ECML PKDD) 2024. The challenge focused on the data-driven predictability of global sea surface temperatures (SSTs), a key factor in climate forecasting, ecosystem management, fisheries management, and climate change monitoring. The challenge involved forecasting SST anomalies (SSTAs) three months in advance using historical data and included a special task of predicting SSTAs nine months ahead for the Baltic Sea. Participants utilized various machine learning approaches to tackle the task, leveraging data from ERA5. This paper discusses the methodologies employed, the results obtained, and the lessons learned, offering insights into the future of climate-related predictive modeling.
2501.05733
TB-Bench: Training and Testing Multi-Modal AI for Understanding Spatio-Temporal Traffic Behaviors from Dashcam Images/Videos
cs.CV
The application of Multi-modal Large Language Models (MLLMs) in Autonomous Driving (AD) faces significant challenges due to their limited training on traffic-specific data and the absence of dedicated benchmarks for spatiotemporal understanding. This study addresses these issues by proposing TB-Bench, a comprehensive benchmark designed to evaluate MLLMs on understanding traffic behaviors across eight perception tasks from ego-centric views. We also introduce vision-language instruction tuning datasets, TB-100k and TB-250k, along with simple yet effective baselines for the tasks. Through extensive experiments, we show that existing MLLMs underperform in these tasks, with even a powerful model like GPT-4o achieving less than 35% accuracy on average. In contrast, when fine-tuned with TB-100k or TB-250k, our baseline models achieve average accuracy up to 85%, significantly enhancing performance on the tasks. Additionally, we demonstrate performance transfer by co-training TB-100k with another traffic dataset, leading to improved performance on the latter. Overall, this study represents a step forward by introducing a comprehensive benchmark, high-quality datasets, and baselines, thus supporting the gradual integration of MLLMs into the perception, prediction, and planning stages of AD.
2501.05735
ELENA: Epigenetic Learning through Evolved Neural Adaptation
cs.NE cs.LG
Despite the success of metaheuristic algorithms in solving complex network optimization problems, they often struggle with adaptation, especially in dynamic or high-dimensional search spaces. Traditional approaches can become stuck in local optima, leading to inefficient exploration and suboptimal solutions. Most of the widely accepted advanced algorithms do well either on highly complex or smaller search spaces due to the lack of adaptation. To address these limitations, we present ELENA (Epigenetic Learning through Evolved Neural Adaptation), a new evolutionary framework that incorporates epigenetic mechanisms to enhance the adaptability of the core evolutionary approach. ELENA leverages compressed representation of learning parameters improved dynamically through epigenetic tags that serve as adaptive memory. Three epigenetic tags (mutation resistance, crossover affinity, and stability score) assist with guiding solution space search, facilitating a more intelligent hypothesis landscape exploration. To assess the framework performance, we conduct experiments on three critical network optimization problems: the Traveling Salesman Problem (TSP), the Vehicle Routing Problem (VRP), and the Maximum Clique Problem (MCP). Experiments indicate that ELENA achieves competitive results, often surpassing state-of-the-art methods on network optimization tasks.
2501.05744
LLVD: LSTM-based Explicit Motion Modeling in Latent Space for Blind Video Denoising
cs.CV cs.LG
Video restoration plays a pivotal role in revitalizing degraded video content by rectifying imperfections caused by various degradations introduced during capturing (sensor noise, motion blur, etc.), saving/sharing (compression, resizing, etc.) and editing. This paper introduces a novel algorithm designed for scenarios where noise is introduced during video capture, aiming to enhance the visual quality of videos by reducing unwanted noise artifacts. We propose the Latent space LSTM Video Denoiser (LLVD), an end-to-end blind denoising model. LLVD uniquely combines spatial and temporal feature extraction, employing Long Short Term Memory (LSTM) within the encoded feature domain. This integration of LSTM layers is crucial for maintaining continuity and minimizing flicker in the restored video. Moreover, processing frames in the encoded feature domain significantly reduces computations, resulting in a very lightweight architecture. LLVD's blind nature makes it versatile for real, in-the-wild denoising scenarios where prior information about noise characteristics is not available. Experiments reveal that LLVD demonstrates excellent performance for both synthetic and captured noise. Specifically, LLVD surpasses the current State-Of-The-Art (SOTA) in RAW denoising by 0.3dB, while also achieving a 59\% reduction in computational complexity.
2501.05745
Covariate Dependent Mixture of Bayesian Networks
stat.ML cs.LG
Learning the structure of Bayesian networks from data provides insights into underlying processes and the causal relationships that generate the data, but its usefulness depends on the homogeneity of the data population, a condition often violated in real-world applications. In such cases, using a single network structure for inference can be misleading, as it may not capture sub-population differences. To address this, we propose a novel approach of modelling a mixture of Bayesian networks where component probabilities depend on individual characteristics. Our method identifies both network structures and demographic predictors of sub-population membership, aiding personalised interventions. We evaluate our method through simulations and a youth mental health case study, demonstrating its potential to improve tailored interventions in health, education, and social policy.
2501.05748
From Bit to Block: Decoding on Erasure Channels
cs.IT math.IT
We provide a general framework for bounding the block error threshold of a linear code $C\subseteq \mathbb{F}_2^N$ over the erasure channel in terms of its bit error threshold. Our approach relies on understanding the minimum support weight of any $r$-dimensional subcode of $C$, for all small values of $r$. As a proof of concept, we use our machinery to obtain a new proof of the celebrated result that Reed-Muller codes achieve capacity on the erasure channel with respect to block error probability.
2501.05749
Bridging Dialects: Translating Standard Bangla to Regional Variants Using Neural Models
cs.CL
The Bangla language includes many regional dialects, adding to its cultural richness. The translation of Bangla Language into regional dialects presents a challenge due to significant variations in vocabulary, pronunciation, and sentence structure across regions like Chittagong, Sylhet, Barishal, Noakhali, and Mymensingh. These dialects, though vital to local identities, lack of representation in technological applications. This study addresses this gap by translating standard Bangla into these dialects using neural machine translation (NMT) models, including BanglaT5, mT5, and mBART50. The work is motivated by the need to preserve linguistic diversity and improve communication among dialect speakers. The models were fine-tuned using the "Vashantor" dataset, containing 32,500 sentences across various dialects, and evaluated through Character Error Rate (CER) and Word Error Rate (WER) metrics. BanglaT5 demonstrated superior performance with a CER of 12.3% and WER of 15.7%, highlighting its effectiveness in capturing dialectal nuances. The outcomes of this research contribute to the development of inclusive language technologies that support regional dialects and promote linguistic diversity.
2501.05750
Semantic Mapping in Indoor Embodied AI -- A Comprehensive Survey and Future Directions
cs.RO cs.CV
Intelligent embodied agents (e.g. robots) need to perform complex semantic tasks in unfamiliar environments. Among many skills that the agents need to possess, building and maintaining a semantic map of the environment is most crucial in long-horizon tasks. A semantic map captures information about the environment in a structured way, allowing the agent to reference it for advanced reasoning throughout the task. While existing surveys in embodied AI focus on general advancements or specific tasks like navigation and manipulation, this paper provides a comprehensive review of semantic map-building approaches in embodied AI, specifically for indoor navigation. We categorize these approaches based on their structural representation (spatial grids, topological graphs, dense point-clouds or hybrid maps) and the type of information they encode (implicit features or explicit environmental data). We also explore the strengths and limitations of the map building techniques, highlight current challenges, and propose future research directions. We identify that the field is moving towards developing open-vocabulary, queryable, task-agnostic map representations, while high memory demands and computational inefficiency still remaining to be open challenges. This survey aims to guide current and future researchers in advancing semantic mapping techniques for embodied AI systems.
2501.05752
Semantic Exploration with Adaptive Gating for Efficient Problem Solving with Language Models
cs.AI cs.CL
Recent advancements in large language models (LLMs) have shown remarkable potential in various complex tasks requiring multi-step reasoning methods like tree search to explore diverse reasoning paths. However, existing methods often suffer from computational inefficiency and redundancy. First, they overlook the diversity of task difficulties, leading to unnecessarily extensive searches even for easy tasks. Second, they neglect the semantics of reasoning paths, resulting in redundant exploration of semantically identical paths. To address these limitations, we propose Semantic Exploration with Adaptive Gating (SEAG), a computationally efficient method. SEAG employs an adaptive gating mechanism that dynamically decides whether to conduct a tree search, based on the confidence level of answers from a preceding simple reasoning method. Furthermore, its tree-based exploration consolidates semantically identical reasoning steps, reducing redundant explorations while maintaining or even improving accuracy. Our extensive experiments demonstrate that SEAG significantly improves accuracy by 4.3% on average while requiring only 31% of computational costs compared to existing tree search-based methods on complex reasoning benchmarks including GSM8K and ARC with diverse language models such as Llama2, Llama3, and Mistral.
2501.05755
CognoSpeak: an automatic, remote assessment of early cognitive decline in real-world conversational speech
cs.SD cs.LG eess.AS
The early signs of cognitive decline are often noticeable in conversational speech, and identifying those signs is crucial in dealing with later and more serious stages of neurodegenerative diseases. Clinical detection is costly and time-consuming and although there has been recent progress in the automatic detection of speech-based cues, those systems are trained on relatively small databases, lacking detailed metadata and demographic information. This paper presents CognoSpeak and its associated data collection efforts. CognoSpeak asks memory-probing long and short-term questions and administers standard cognitive tasks such as verbal and semantic fluency and picture description using a virtual agent on a mobile or web platform. In addition, it collects multimodal data such as audio and video along with a rich set of metadata from primary and secondary care, memory clinics and remote settings like people's homes. Here, we present results from 126 subjects whose audio was manually transcribed. Several classic classifiers, as well as large language model-based classifiers, have been investigated and evaluated across the different types of prompts. We demonstrate a high level of performance; in particular, we achieved an F1-score of 0.873 using a DistilBERT model to discriminate people with cognitive impairment (dementia and people with mild cognitive impairment (MCI)) from healthy volunteers using the memory responses, fluency tasks and cookie theft picture description. CognoSpeak is an automatic, remote, low-cost, repeatable, non-invasive and less stressful alternative to existing clinical cognitive assessments.
2501.05757
Locality-aware Gaussian Compression for Fast and High-quality Rendering
cs.CV
We present LocoGS, a locality-aware 3D Gaussian Splatting (3DGS) framework that exploits the spatial coherence of 3D Gaussians for compact modeling of volumetric scenes. To this end, we first analyze the local coherence of 3D Gaussian attributes, and propose a novel locality-aware 3D Gaussian representation that effectively encodes locally-coherent Gaussian attributes using a neural field representation with a minimal storage requirement. On top of the novel representation, LocoGS is carefully designed with additional components such as dense initialization, an adaptive spherical harmonics bandwidth scheme and different encoding schemes for different Gaussian attributes to maximize compression performance. Experimental results demonstrate that our approach outperforms the rendering quality of existing compact Gaussian representations for representative real-world 3D datasets while achieving from 54.6$\times$ to 96.6$\times$ compressed storage size and from 2.1$\times$ to 2.4$\times$ rendering speed than 3DGS. Even our approach also demonstrates an averaged 2.4$\times$ higher rendering speed than the state-of-the-art compression method with comparable compression performance.
2501.05762
Development and Comparison of Model-Based and Data-Driven Approaches for the Prediction of the Mechanical Properties of Lattice Structures
cond-mat.soft cs.CE cs.LG physics.comp-ph
Lattice structures have great potential for several application fields ranging from medical and tissue engineering to aeronautical one. Their development is further speeded up by the continuing advances in additive manufacturing technologies that allow to overcome issues typical of standard processes and to propose tailored designs. However, the design of lattice structures is still challenging since their properties are considerably affected by numerous factors. The present paper aims to propose, discuss, and compare various modeling approaches to describe, understand, and predict the correlations between the mechanical properties and the void volume fraction of different types of lattice structures fabricated by fused deposition modeling 3D printing. Particularly, four approaches are proposed: (i) a simplified analytical model; (ii) a semi-empirical model combining analytical equations with experimental correction factors; (iii) an artificial neural network trained on experimental data; (iv) numerical simulations by finite element analyses. The comparison among the various approaches, and with experimental data, allows to identify the performances, advantages, and disadvantages of each approach, thus giving important guidelines for choosing the right design methodology based on the needs and available data.
2501.05763
StarGen: A Spatiotemporal Autoregression Framework with Video Diffusion Model for Scalable and Controllable Scene Generation
cs.CV
Recent advances in large reconstruction and generative models have significantly improved scene reconstruction and novel view generation. However, due to compute limitations, each inference with these large models is confined to a small area, making long-range consistent scene generation challenging. To address this, we propose StarGen, a novel framework that employs a pre-trained video diffusion model in an autoregressive manner for long-range scene generation. The generation of each video clip is conditioned on the 3D warping of spatially adjacent images and the temporally overlapping image from previously generated clips, improving spatiotemporal consistency in long-range scene generation with precise pose control. The spatiotemporal condition is compatible with various input conditions, facilitating diverse tasks, including sparse view interpolation, perpetual view generation, and layout-conditioned city generation. Quantitative and qualitative evaluations demonstrate StarGen's superior scalability, fidelity, and pose accuracy compared to state-of-the-art methods.
2501.05764
Controlling Large Language Models Through Concept Activation Vectors
cs.CL
As large language models (LLMs) are widely deployed across various domains, the ability to control their generated outputs has become more critical. This control involves aligning LLMs outputs with human values and ethical principles or customizing LLMs on specific topics or styles for individual users. Existing controlled generation methods either require significant computational resources and extensive trial-and-error or provide coarse-grained control. In this paper, we propose Generation with Concept Activation Vector (GCAV), a lightweight model control framework that ensures accurate control without requiring resource-extensive fine-tuning. Specifically, GCAV first trains a concept activation vector for specified concepts to be controlled, such as toxicity. During inference, GCAV steers the concept vector in LLMs, for example, by removing the toxicity concept vector from the activation layers. Control experiments from different perspectives, including toxicity reduction, sentiment control, linguistic style, and topic control, demonstrate that our framework achieves state-of-the-art performance with granular control, allowing for fine-grained adjustments of both the steering layers and the steering magnitudes for individual samples.
2501.05765
Deontic Temporal Logic for Formal Verification of AI Ethics
cs.AI cs.LO
Ensuring ethical behavior in Artificial Intelligence (AI) systems amidst their increasing ubiquity and influence is a major concern the world over. The use of formal methods in AI ethics is a possible crucial approach for specifying and verifying the ethical behavior of AI systems. This paper proposes a formalization based on deontic logic to define and evaluate the ethical behavior of AI systems, focusing on system-level specifications, contributing to this important goal. It introduces axioms and theorems to capture ethical requirements related to fairness and explainability. The formalization incorporates temporal operators to reason about the ethical behavior of AI systems over time. The authors evaluate the effectiveness of this formalization by assessing the ethics of the real-world COMPAS and loan prediction AI systems. Various ethical properties of the COMPAS and loan prediction systems are encoded using deontic logical formulas, allowing the use of an automated theorem prover to verify whether these systems satisfy the defined properties. The formal verification reveals that both systems fail to fulfill certain key ethical properties related to fairness and non-discrimination, demonstrating the effectiveness of the proposed formalization in identifying potential ethical issues in real-world AI applications.
2501.05767
Migician: Revealing the Magic of Free-Form Multi-Image Grounding in Multimodal Large Language Models
cs.CL cs.AI cs.CV
The recent advancement of Multimodal Large Language Models (MLLMs) has significantly improved their fine-grained perception of single images and general comprehension across multiple images. However, existing MLLMs still face challenges in achieving precise grounding in complex multi-image scenarios. To address this, we first explore a Chain-of-Thought (CoT) framework that integrates single-image grounding with multi-image comprehension. While partially effective, it remains unstable and struggles to capture abstract visual information due to its non-end-to-end nature. Therefore, we introduce Migician, the first multi-image grounding model capable of performing free-form and accurate grounding across multiple images. To support this, we present the MGrounding-630k dataset, which comprises data for several multi-image grounding tasks derived from existing datasets, along with newly generated free-form grounding instruction-following data. Furthermore, we propose MIG-Bench, a comprehensive benchmark specifically designed for evaluating multi-image grounding capabilities. Experimental results demonstrate that our model achieves significantly superior multi-image grounding capabilities, outperforming the best existing MLLMs by 24.94% and even surpassing much larger 70B models. Our code, model, dataset, and benchmark are fully open-sourced at https://migician-vg.github.io/.
2501.05768
Halal or Not: Knowledge Graph Completion for Predicting Cultural Appropriateness of Daily Products
cs.LG cs.AI
The growing demand for halal cosmetic products has exposed significant challenges, especially in Muslim-majority countries. Recently, various machine learning-based strategies, e.g., image-based methods, have shown remarkable success in predicting the halal status of cosmetics. However, these methods mainly focus on analyzing the discrete and specific ingredients within separate cosmetics, which ignore the high-order and complex relations between cosmetics and ingredients. To address this problem, we propose a halal cosmetic recommendation framework, namely HaCKG, that leverages a knowledge graph of cosmetics and their ingredients to explicitly model and capture the relationships between cosmetics and their components. By representing cosmetics and ingredients as entities within the knowledge graph, HaCKG effectively learns the high-order and complex relations between entities, offering a robust method for predicting halal status. Specifically, we first construct a cosmetic knowledge graph representing the relations between various cosmetics, ingredients, and their properties. We then propose a pre-trained relational graph attention network model with residual connections to learn the structural relation between entities in the knowledge graph. The pre-trained model is then fine-tuned on downstream cosmetic data to predict halal status. Extensive experiments on the cosmetic dataset over halal prediction tasks demonstrate the superiority of our model over state-of-the-art baselines.
2501.05769
Conditional Diffusion Model for Electrical Impedance Tomography
cs.CV
Electrical impedance tomography (EIT) is a non-invasive imaging technique, which has been widely used in the fields of industrial inspection, medical monitoring and tactile sensing. However, due to the inherent non-linearity and ill-conditioned nature of the EIT inverse problem, the reconstructed image is highly sensitive to the measured data, and random noise artifacts often appear in the reconstructed image, which greatly limits the application of EIT. To address this issue, a conditional diffusion model with voltage consistency (CDMVC) is proposed in this study. The method consists of a pre-imaging module, a conditional diffusion model for reconstruction, a forward voltage constraint network and a scheme of voltage consistency constraint during sampling process. The pre-imaging module is employed to generate the initial reconstruction. This serves as a condition for training the conditional diffusion model. Finally, based on the forward voltage constraint network, a voltage consistency constraint is implemented in the sampling phase to incorporate forward information of EIT, thereby enhancing imaging quality. A more complete dataset, including both common and complex concave shapes, is generated. The proposed method is validated using both simulation and physical experiments. Experimental results demonstrate that our method can significantly improves the quality of reconstructed images. In addition, experimental results also demonstrate that our method has good robustness and generalization performance.
2501.05770
Path Planning for Multi-Copter UAV Formation Employing a Generalized Particle Swarm Optimization
cs.RO cs.SY eess.SY
The paper investigates the problem of path planning techniques for multi-copter uncrewed aerial vehicles (UAV) cooperation in a formation shape to examine surrounding surfaces. We first describe the problem as a joint objective cost for planning a path of the formation centroid working in a complicated space. The path planning algorithm, named the generalized particle swarm optimization algorithm, is then presented to construct an optimal, flyable path while avoiding obstacles and ensuring the flying mission requirements. A path-development scheme is then incorporated to generate a relevant path for each drone to maintain its position in the formation configuration. Simulation, comparison, and experiments have been conducted to verify the proposed approach. Results show the feasibility of the proposed path-planning algorithm with GEPSO.
2501.05772
rmlnomogram: An R package to construct an explainable nomogram for any machine learning algorithms
cs.LG stat.ML
Background: Current nomogram can only be created for regression algorithm. Providing nomogram for any machine learning (ML) algorithms may accelerate model deployment in clinical settings or improve model availability. We developed an R package and web application to construct nomogram with model explainability of any ML algorithms. Methods: We formulated a function to transform an ML prediction model into a nomogram, requiring datasets with: (1) all possible combinations of predictor values; (2) the corresponding outputs of the model; and (3) the corresponding explainability values for each predictor (optional). Web application was also created. Results: Our R package could create 5 types of nomograms for categorical predictors and binary outcome without probability (1), categorical predictors and binary outcome with probability (2) or continuous outcome (3), and categorical with single numerical predictors and binary outcome with probability (4) or continuous outcome (5). Respectively, the first and remaining types optimally allowed maximum 15 and 5 predictors with maximum 3,200 combinations. Web application is provided with such limits. The explainability values were possible for types 2 to 5. Conclusions: Our R package and web application could construct nomogram with model explainability of any ML algorithms using a fair number of predictors.
2501.05775
STHFL: Spatio-Temporal Heterogeneous Federated Learning
cs.LG cs.DC
Federated learning is a new framework that protects data privacy and allows multiple devices to cooperate in training machine learning models. Previous studies have proposed multiple approaches to eliminate the challenges posed by non-iid data and inter-domain heterogeneity issues. However, they ignore the \textbf{spatio-temporal} heterogeneity formed by different data distributions of increasing task data in the intra-domain. Moreover, the global data is generally a long-tailed distribution rather than assuming the global data is balanced in practical applications. To tackle the \textbf{spatio-temporal} dilemma, we propose a novel setting named \textbf{Spatio-Temporal Heterogeneity} Federated Learning (STHFL). Specially, the Global-Local Dynamic Prototype (GLDP) framework is designed for STHFL. In GLDP, the model in each client contains personalized layers which can dynamically adapt to different data distributions. For long-tailed data distribution, global prototypes are served as complementary knowledge for the training on classes with few samples in clients without leaking privacy. As tasks increase in clients, the knowledge of local prototypes generated in previous tasks guides for training in the current task to solve catastrophic forgetting. Meanwhile, the global-local prototypes are updated through the moving average method after training local prototypes in clients. Finally, we evaluate the effectiveness of GLDP, which achieves remarkable results compared to state-of-the-art methods in STHFL scenarios.
2501.05777
StructSR: Refuse Spurious Details in Real-World Image Super-Resolution
cs.CV
Diffusion-based models have shown great promise in real-world image super-resolution (Real-ISR), but often generate content with structural errors and spurious texture details due to the empirical priors and illusions of these models. To address this issue, we introduce StructSR, a simple, effective, and plug-and-play method that enhances structural fidelity and suppresses spurious details for diffusion-based Real-ISR. StructSR operates without the need for additional fine-tuning, external model priors, or high-level semantic knowledge. At its core is the Structure-Aware Screening (SAS) mechanism, which identifies the image with the highest structural similarity to the low-resolution (LR) input in the early inference stage, allowing us to leverage it as a historical structure knowledge to suppress the generation of spurious details. By intervening in the diffusion inference process, StructSR seamlessly integrates with existing diffusion-based Real-ISR models. Our experimental results demonstrate that StructSR significantly improves the fidelity of structure and texture, improving the PSNR and SSIM metrics by an average of 5.27% and 9.36% on a synthetic dataset (DIV2K-Val) and 4.13% and 8.64% on two real-world datasets (RealSR and DRealSR) when integrated with four state-of-the-art diffusion-based Real-ISR methods.
2501.05778
Formally Verified Neural Lyapunov Function for Incremental Input-to-State Stability of Unknown Systems
eess.SY cs.SY
This work presents an approach to synthesize a Lyapunov-like function to ensure incrementally input-to-state stability ($\delta$-ISS) property for an unknown discrete-time system. To deal with challenges posed by unknown system dynamics, we parameterize the Lyapunov-like function as a neural network, which we train using the data samples collected from the unknown system along with appropriately designed loss functions. We propose a validity condition to test the obtained function and incorporate it into the training framework to ensure provable correctness at the end of the training. Finally, the usefulness of the proposed technique is proved using two case studies: a scalar non-linear dynamical system and a permanent magnet DC motor.
2501.05780
Multi-layer RIS on Edge: Communication, Computation and Wireless Power Transfer
cs.IT math.IT
The rapid expansion of Internet of Things (IoT) and its integration into various applications highlight the need for advanced communication, computation, and energy transfer techniques. However, the traditional hardware-based evolution of communication systems faces challenges due to excessive power consumption and prohibitive hardware cost. With the rapid advancement of reconfigurable intelligent surface (RIS), a new approach by parallel stacking a series of RIS, i.e., multi-layer RIS, has been proposed. Benefiting from the characteristics of scalability, passivity, low cost, and enhanced computation capability, multi-layer RIS is a promising technology for future massive IoT scenarios. Thus, this article proposes a multi-layer RIS-based universal paradigm at the network edge, enabling three functions, i.e., multiple-input multiple-output (MIMO) communication, computation, and wireless power transfer (WPT). Starting by picturing the possible applications of multi-layer RIS, we explore the potential signal transmission links, energy transmission links, and computation processes in IoT scenarios, showing its ability to handle on-edge IoT tasks and associated green challenges. Then, these three key functions are analyzed respectively in detail, showing the advantages of the proposed scheme, compared with the traditional hardware-based scheme. To facilitate the implementation of this new paradigm into reality, we list the dominant future research directions at last, such as inter-layer channel modeling, resource allocation and scheduling, channel estimation, and edge training. It is anticipated that multi-layer RIS will contribute to more energy-efficient wireless networks in the future by introducing a revolutionary paradigm shift to an all-wave-based approach.
2501.05783
UV-Attack: Physical-World Adversarial Attacks for Person Detection via Dynamic-NeRF-based UV Mapping
cs.CV cs.AI
In recent research, adversarial attacks on person detectors using patches or static 3D model-based texture modifications have struggled with low success rates due to the flexible nature of human movement. Modeling the 3D deformations caused by various actions has been a major challenge. Fortunately, advancements in Neural Radiance Fields (NeRF) for dynamic human modeling offer new possibilities. In this paper, we introduce UV-Attack, a groundbreaking approach that achieves high success rates even with extensive and unseen human actions. We address the challenge above by leveraging dynamic-NeRF-based UV mapping. UV-Attack can generate human images across diverse actions and viewpoints, and even create novel actions by sampling from the SMPL parameter space. While dynamic NeRF models are capable of modeling human bodies, modifying clothing textures is challenging because they are embedded in neural network parameters. To tackle this, UV-Attack generates UV maps instead of RGB images and modifies the texture stacks. This approach enables real-time texture edits and makes the attack more practical. We also propose a novel Expectation over Pose Transformation loss (EoPT) to improve the evasion success rate on unseen poses and views. Our experiments show that UV-Attack achieves a 92.75% attack success rate against the FastRCNN model across varied poses in dynamic video settings, significantly outperforming the state-of-the-art AdvCamou attack, which only had a 28.50% ASR. Moreover, we achieve 49.5% ASR on the latest YOLOv8 detector in black-box settings. This work highlights the potential of dynamic NeRF-based UV mapping for creating more effective adversarial attacks on person detectors, addressing key challenges in modeling human movement and texture modification.
2501.05786
Cryptanalysis of Cancelable Biometrics Vault
cs.CR cs.CV
Cancelable Biometrics (CB) stands for a range of biometric transformation schemes combining biometrics with user specific tokens to generate secure templates. Required properties are the irreversibility, unlikability and recognition accuracy of templates while making their revocation possible. In biometrics, a key-binding scheme is used for protecting a cryptographic key using a biometric data. The key can be recomputed only if a correct biometric data is acquired during authentication. Applications of key-binding schemes are typically disk encryption, where the cryptographic key is used to encrypt and decrypt the disk. In this paper, we cryptanalyze a recent key-binding scheme, called Cancelable Biometrics Vault (CBV) based on cancelable biometrics. More precisely, the introduced cancelable transformation, called BioEncoding scheme, for instantiating the CBV framework is attacked in terms of reversibility and linkability of templates. Subsequently, our linkability attack enables to recover the key in the vault without additional assumptions. Our cryptanalysis introduces a new perspective by uncovering the CBV scheme's revocability and linkability vulnerabilities, which were not previously identified in comparable biometric-based key-binding schemes.
2501.05787
MARS6: A Small and Robust Hierarchical-Codec Text-to-Speech Model
eess.AS cs.CL
Codec-based text-to-speech (TTS) models have shown impressive quality with zero-shot voice cloning abilities. However, they often struggle with more expressive references or complex text inputs. We present MARS6, a robust encoder-decoder transformer for rapid, expressive TTS. MARS6 is built on recent improvements in spoken language modelling. Utilizing a hierarchical setup for its decoder, new speech tokens are processed at a rate of only 12 Hz, enabling efficient modelling of long-form text while retaining reconstruction quality. We combine several recent training and inference techniques to reduce repetitive generation and improve output stability and quality. This enables the 70M-parameter MARS6 to achieve similar performance to models many times larger. We show this in objective and subjective evaluations, comparing TTS output quality and reference speaker cloning ability. Project page: https://camb-ai.github.io/mars6-turbo/
2501.05790
Understanding Impact of Human Feedback via Influence Functions
cs.AI cs.HC cs.LG
In Reinforcement Learning from Human Feedback (RLHF), it is crucial to learn suitable reward models from human feedback to align large language models (LLMs) with human intentions. However, human feedback can often be noisy, inconsistent, or biased, especially when evaluating complex responses. Such feedback can lead to misaligned reward signals, potentially causing unintended side effects during the RLHF process. To address these challenges, we explore the use of influence functions to measure the impact of human feedback on the performance of reward models. We propose a compute-efficient approximation method that enables the application of influence functions to LLM-based reward models and large-scale preference datasets. In our experiments, we demonstrate two key applications of influence functions: (1) detecting common forms of labeler bias in human feedback datasets and (2) guiding labelers to refine their strategies to align more closely with expert feedback. By quantifying the impact of human feedback on reward models, we believe that influence functions can enhance feedback interpretability and contribute to scalable oversight in RLHF, helping labelers provide more accurate and consistent feedback. Source code is available at https://github.com/mintaywon/IF_RLHF
2501.05795
Robust Counterfactual Explanations under Model Multiplicity Using Multi-Objective Optimization
cs.LG cs.AI
In recent years, explainability in machine learning has gained importance. In this context, counterfactual explanation (CE), which is an explanation method that uses examples, has attracted attention. However, it has been pointed out that CE is not robust when there are multiple machine-learning models with similar accuracy. These problems are important when using machine learning to make safe decisions. In this paper, we propose robust CEs that introduce a new viewpoint -- Pareto improvement -- and a method that uses multi-objective optimization to generate it. To evaluate the proposed method, we conducted experiments using both simulated and real data. The results demonstrate that the proposed method is both robust and practical. This study highlights the potential of ensuring robustness in decision-making by applying the concept of social welfare. We believe that this research can serve as a valuable foundation for various fields, including explainability in machine learning, decision-making, and action planning based on machine learning.
2501.05803
Test-time Alignment of Diffusion Models without Reward Over-optimization
cs.LG cs.AI cs.CV math.ST stat.TH
Diffusion models excel in generative tasks, but aligning them with specific objectives while maintaining their versatility remains challenging. Existing fine-tuning methods often suffer from reward over-optimization, while approximate guidance approaches fail to optimize target rewards effectively. Addressing these limitations, we propose a training-free, test-time method based on Sequential Monte Carlo (SMC) to sample from the reward-aligned target distribution. Our approach, tailored for diffusion sampling and incorporating tempering techniques, achieves comparable or superior target rewards to fine-tuning methods while preserving diversity and cross-reward generalization. We demonstrate its effectiveness in single-reward optimization, multi-objective scenarios, and online black-box optimization. This work offers a robust solution for aligning diffusion models with diverse downstream objectives without compromising their general capabilities. Code is available at https://github.com/krafton-ai/DAS.
2501.05808
Real-Time Integrated Dispatching and Idle Fleet Steering with Deep Reinforcement Learning for A Meal Delivery Platform
eess.SY cs.AI cs.SY
To achieve high service quality and profitability, meal delivery platforms like Uber Eats and Grubhub must strategically operate their fleets to ensure timely deliveries for current orders while mitigating the consequential impacts of suboptimal decisions that leads to courier understaffing in the future. This study set out to solve the real-time order dispatching and idle courier steering problems for a meal delivery platform by proposing a reinforcement learning (RL)-based strategic dual-control framework. To address the inherent sequential nature of these problems, we model both order dispatching and courier steering as Markov Decision Processes. Trained via a deep reinforcement learning (DRL) framework, we obtain strategic policies by leveraging the explicitly predicted demands as part of the inputs. In our dual-control framework, the dispatching and steering policies are iteratively trained in an integrated manner. These forward-looking policies can be executed in real-time and provide decisions while jointly considering the impacts on local and network levels. To enhance dispatching fairness, we propose convolutional deep Q networks to construct fair courier embeddings. To simultaneously rebalance the supply and demand within the service network, we propose to utilize mean-field approximated supply-demand knowledge to reallocate idle couriers at the local level. Utilizing the policies generated by the RL-based strategic dual-control framework, we find the delivery efficiency and fairness of workload distribution among couriers have been improved, and under-supplied conditions have been alleviated within the service network. Our study sheds light on designing an RL-based framework to enable forward-looking real-time operations for meal delivery platforms and other on-demand services.
2501.05809
AdaPRL: Adaptive Pairwise Regression Learning with Uncertainty Estimation for Universal Regression Tasks
cs.LG
Current deep regression models usually learn in a point-wise way that treats each sample as an independent input, neglecting the relative ordering among different data. Consequently, the regression model could neglect the data's interrelationships, potentially resulting in suboptimal performance. Moreover, the existence of aleatoric uncertainty in the training data may drive the model to capture non-generalizable patterns, contributing to increased overfitting. To address these issues, we propose a novel adaptive pairwise learning framework for regression tasks (AdaPRL) which leverages the relative differences between data points and integrates with deep probabilistic models to quantify the uncertainty associated with the predictions. Additionally, we adapt AdaPRL for applications in multi-task learning and multivariate time series forecasting. Extensive experiments with several real-world regression datasets including recommendation systems, age prediction, time series forecasting, natural language understanding, finance, and industry datasets show that AdaPRL is compatible with different backbone networks in various tasks and achieves state-of-the-art performance on the vast majority of tasks without extra inference cost, highlighting its notable potential including enhancing prediction accuracy and ranking ability, increasing generalization capability, improving robustness to noisy data, improving resilience to reduced data, and enhancing interpretability. Experiments also show that AdaPRL can be seamlessly incorporated into recently proposed regression frameworks to gain performance improvement.
2501.05813
Social web and Wikipedia: an opportunity to rethink the links between sources' credibility, trust and authority
cs.IR cs.CY cs.SI
The Web and its main tools (Google, Wikipedia, Facebook, Twitter) deeply raise and renew fundamental questions, that everyone asks almost every day: Is this information or content true? Can I trust this author or source? These questions are not new, they have been the same with books, newspapers, broadcasting and television, and, more fundamentally, in every human interpersonal communication. This paper is focused on two scientific problems on this issue. The first one is theoretical: to address this issue, many concepts have been used in library and information sciences, communication and psychology. The links between these concepts are not clear: sometimes two concepts are considered as synonymous, sometimes as very different. The second one is historical: sources like Wikipedia deeply challenge the epistemic evaluation of information sources, compared to previous modes of information production. This paper proposes an integrated and simple model considering the relation between a user, a document and an author as human communication. It reduces the problem to three concepts: credibility as a characteristic granted to information depending on its truth-value; trust as the ability to produce credible information; authority when the power to influence of an author is accepted, i.e., when readers accept that the source can modify their opinion, knowledge and decisions. The model describes also two kinds of relationships between the three concepts: an upward link and a downward link. The model is confronted with findings of empirical research on Wikipedia in particular.
2501.05815
Enhanced sampled-data model predictive control via nonlinear lifting
eess.SY cs.SY
This paper introduces a novel nonlinear model predictive control (NMPC) framework that incorporates a lifting technique to enhance control performance for nonlinear systems. While the lifting technique has been widely employed in linear systems to capture intersample behaviour, their application to nonlinear systems remains unexplored. We address this gap by formulating an NMPC scheme that combines fast-sample fast-hold (FSFH) approximations and numerical methods to approximate system dynamics and cost functions. The proposed approach is validated through two case studies: the Van der Pol oscillator and the inverted pendulum on a cart. Simulation results demonstrate that the lifted NMPC outperforms conventional NMPC in terms of reduced settling time and improved control accuracy. These findings underscore the potential of the lifting-based NMPC for efficient control of nonlinear systems, offering a practical solution for real-time applications.
2501.05816
IndoNLP 2025: Shared Task on Real-Time Reverse Transliteration for Romanized Indo-Aryan languages
cs.CL
The paper overviews the shared task on Real-Time Reverse Transliteration for Romanized Indo-Aryan languages. It focuses on the reverse transliteration of low-resourced languages in the Indo-Aryan family to their native scripts. Typing Romanized Indo-Aryan languages using ad-hoc transliterals and achieving accurate native scripts are complex and often inaccurate processes with the current keyboard systems. This task aims to introduce and evaluate a real-time reverse transliterator that converts Romanized Indo-Aryan languages to their native scripts, improving the typing experience for users. Out of 11 registered teams, four teams participated in the final evaluation phase with transliteration models for Sinhala, Hindi and Malayalam. These proposed solutions not only solve the issue of ad-hoc transliteration but also empower low-resource language usability in the digital arena.
2501.05819
Diffusion Models for Smarter UAVs: Decision-Making and Modeling
cs.LG cs.AI
Unmanned Aerial Vehicles (UAVs) are increasingly adopted in modern communication networks. However, challenges in decision-making and digital modeling continue to impede their rapid advancement. Reinforcement Learning (RL) algorithms face limitations such as low sample efficiency and limited data versatility, further magnified in UAV communication scenarios. Moreover, Digital Twin (DT) modeling introduces substantial decision-making and data management complexities. RL models, often integrated into DT frameworks, require extensive training data to achieve accurate predictions. In contrast to traditional approaches that focus on class boundaries, Diffusion Models (DMs), a new class of generative AI, learn the underlying probability distribution from the training data and can generate trustworthy new patterns based on this learned distribution. This paper explores the integration of DMs with RL and DT to effectively address these challenges. By combining the data generation capabilities of DMs with the decision-making framework of RL and the modeling accuracy of DT, the integration improves the adaptability and real-time performance of UAV communication. Moreover, the study shows how DMs can alleviate data scarcity, improve policy networks, and optimize dynamic modeling, providing a robust solution for complex UAV communication scenarios.
2501.05823
PersonaHOI: Effortlessly Improving Personalized Face with Human-Object Interaction Generation
cs.CV
We introduce PersonaHOI, a training- and tuning-free framework that fuses a general StableDiffusion model with a personalized face diffusion (PFD) model to generate identity-consistent human-object interaction (HOI) images. While existing PFD models have advanced significantly, they often overemphasize facial features at the expense of full-body coherence, PersonaHOI introduces an additional StableDiffusion (SD) branch guided by HOI-oriented text inputs. By incorporating cross-attention constraints in the PFD branch and spatial merging at both latent and residual levels, PersonaHOI preserves personalized facial details while ensuring interactive non-facial regions. Experiments, validated by a novel interaction alignment metric, demonstrate the superior realism and scalability of PersonaHOI, establishing a new standard for practical personalized face with HOI generation. Our code will be available at https://github.com/JoyHuYY1412/PersonaHOI
2501.05826
AI-Driven Diabetic Retinopathy Screening: Multicentric Validation of AIDRSS in India
eess.IV cs.AI cs.CV
Purpose: Diabetic retinopathy (DR) is a major cause of vision loss, particularly in India, where access to retina specialists is limited in rural areas. This study aims to evaluate the Artificial Intelligence-based Diabetic Retinopathy Screening System (AIDRSS) for DR detection and prevalence assessment, addressing the growing need for scalable, automated screening solutions in resource-limited settings. Approach: A multicentric, cross-sectional study was conducted in Kolkata, India, involving 5,029 participants and 10,058 macula-centric retinal fundus images. The AIDRSS employed a deep learning algorithm with 50 million trainable parameters, integrated with Contrast Limited Adaptive Histogram Equalization (CLAHE) preprocessing for enhanced image quality. DR was graded using the International Clinical Diabetic Retinopathy (ICDR) Scale, categorizing disease into five stages (DR0 to DR4). Statistical metrics including sensitivity, specificity, and prevalence rates were evaluated against expert retina specialist assessments. Results: The prevalence of DR in the general population was 13.7%, rising to 38.2% among individuals with elevated random blood glucose levels. The AIDRSS achieved an overall sensitivity of 92%, specificity of 88%, and 100% sensitivity for detecting referable DR (DR3 and DR4). These results demonstrate the system's robust performance in accurately identifying and grading DR in a diverse population. Conclusions: AIDRSS provides a reliable, scalable solution for early DR detection in resource-constrained environments. Its integration of advanced AI techniques ensures high diagnostic accuracy, with potential to significantly reduce the burden of diabetes-related vision loss in underserved regions.
2501.05828
UltraRay: Full-Path Ray Tracing for Enhancing Realism in Ultrasound Simulation
cs.CV cs.GR
Traditional ultrasound simulators solve the wave equation to model pressure distribution fields, achieving high accuracy but requiring significant computational time and resources. To address this, ray tracing approaches have been introduced, modeling wave propagation as rays interacting with boundaries and scatterers. However, existing models simplify ray propagation, generating echoes at interaction points without considering return paths to the sensor. This can result in unrealistic artifacts and necessitates careful scene tuning for plausible results. We propose a novel ultrasound simulation pipeline that utilizes a ray tracing algorithm to generate echo data, tracing each ray from the transducer through the scene and back to the sensor. To replicate advanced ultrasound imaging, we introduce a ray emission scheme optimized for plane wave imaging, incorporating delay and steering capabilities. Furthermore, we integrate a standard signal processing pipeline to simulate end-to-end ultrasound image formation. We showcase the efficacy of the proposed pipeline by modeling synthetic scenes featuring highly reflective objects, such as bones. In doing so, our proposed approach, UltraRay, not only enhances the overall visual quality but also improves the realism of the simulated images by accurately capturing secondary reflections and reducing unnatural artifacts. By building on top of a differentiable framework, the proposed pipeline lays the groundwork for a fast and differentiable ultrasound simulation tool necessary for gradient-based optimization, enabling advanced ultrasound beamforming strategies, neural network integration, and accurate inverse scene reconstruction.
2501.05835
Fine-tuning is Not Fine: Mitigating Backdoor Attacks in GNNs with Limited Clean Data
cs.LG cs.CR
Graph Neural Networks (GNNs) have achieved remarkable performance through their message-passing mechanism. However, recent studies have highlighted the vulnerability of GNNs to backdoor attacks, which can lead the model to misclassify graphs with attached triggers as the target class. The effectiveness of recent promising defense techniques, such as fine-tuning or distillation, is heavily contingent on having comprehensive knowledge of the sufficient training dataset. Empirical studies have shown that fine-tuning methods require a clean dataset of 20% to reduce attack accuracy to below 25%, while distillation methods require a clean dataset of 15%. However, obtaining such a large amount of clean data is commonly impractical. In this paper, we propose a practical backdoor mitigation framework, denoted as GRAPHNAD, which can capture high-quality intermediate-layer representations in GNNs to enhance the distillation process with limited clean data. To achieve this, we address the following key questions: How to identify the appropriate attention representations in graphs for distillation? How to enhance distillation with limited data? By adopting the graph attention transfer method, GRAPHNAD can effectively align the intermediate-layer attention representations of the backdoored model with that of the teacher model, forcing the backdoor neurons to transform into benign ones. Besides, we extract the relation maps from intermediate-layer transformation and enforce the relation maps of the backdoored model to be consistent with that of the teacher model, thereby ensuring model accuracy while further reducing the influence of backdoors. Extensive experimental results show that by fine-tuning a teacher model with only 3% of the clean data, GRAPHNAD can reduce the attack success rate to below 5%.
2501.05839
Poetry in Pixels: Prompt Tuning for Poem Image Generation via Diffusion Models
cs.CV
The task of text-to-image generation has encountered significant challenges when applied to literary works, especially poetry. Poems are a distinct form of literature, with meanings that frequently transcend beyond the literal words. To address this shortcoming, we propose a PoemToPixel framework designed to generate images that visually represent the inherent meanings of poems. Our approach incorporates the concept of prompt tuning in our image generation framework to ensure that the resulting images closely align with the poetic content. In addition, we propose the PoeKey algorithm, which extracts three key elements in the form of emotions, visual elements, and themes from poems to form instructions which are subsequently provided to a diffusion model for generating corresponding images. Furthermore, to expand the diversity of the poetry dataset across different genres and ages, we introduce MiniPo, a novel multimodal dataset comprising 1001 children's poems and images. Leveraging this dataset alongside PoemSum, we conducted both quantitative and qualitative evaluations of image generation using our PoemToPixel framework. This paper demonstrates the effectiveness of our approach and offers a fresh perspective on generating images from literary sources.
2501.05842
Orthogonal projection-based regularization for efficient model augmentation
cs.LG cs.SY eess.SY
Deep-learning-based nonlinear system identification has shown the ability to produce reliable and highly accurate models in practice. However, these black-box models lack physical interpretability, and often a considerable part of the learning effort is spent on capturing already expected/known behavior due to first-principles-based understanding of some aspects of the system. A potential solution is to integrate prior physical knowledge directly into the model structure, combining the strengths of physics-based modeling and deep-learning-based identification. The most common approach is to use an additive model augmentation structure, where the physics-based and the machine-learning (ML) components are connected in parallel. However, such models are overparametrized, training them is challenging, potentially causing the physics-based part to lose interpretability. To overcome this challenge, this paper proposes an orthogonal projection-based regularization technique to enhance parameter learning, convergence, and even model accuracy in learning-based augmentation of nonlinear baseline models.
2501.05844
"Cause" is Mechanistic Narrative within Scientific Domains: An Ordinary Language Philosophical Critique of "Causal Machine Learning"
cs.LG
Causal Learning has emerged as a major theme of research in statistics and machine learning in recent years, promising specific computational techniques to apply to datasets that reveal the true nature of cause and effect in a number of important domains. In this paper we consider the epistemology of recognizing true cause and effect phenomena. We apply the Ordinary Language method of engaging on the customary use of the word 'cause' to investigate valid semantics of reasoning about cause and effect. We recognize that the grammars of cause and effect are fundamentally distinct in form across scientific domains, yet they maintain a consistent and central function. This function can best be described as the mechanism underlying fundamental forces of influence as considered prominent in the respective scientific domain. We demarcate 1) physics and engineering as domains wherein mathematical models are sufficient to comprehensively describe causality, 2) biology as introducing challenges of emergence while providing opportunities for showing consistent mechanisms across scale, and 3) the social sciences as introducing grander difficulties for establishing models of low prediction error but providing, through Hermeneutics, the potential for findings that are still instrumentally useful to individuals. We posit that definitive causal claims regarding a given phenomenon (writ large) can only come through an agglomeration of consistent evidence across multiple domains. This presents important methodological questions as far as harmonizing between language games and emergence across scales. Given the role of epistemic hubris in the contemporary crisis of credibility in the sciences, exercising greater caution as far as communicating precision as to the real degree of certainty certain evidence provides for rich collections of open problems in optimizing integration of different findings.
2501.05845
Annealing Machine-assisted Learning of Graph Neural Network for Combinatorial Optimization
cs.AI cs.LG
While Annealing Machines (AM) have shown increasing capabilities in solving complex combinatorial problems, positioning themselves as a more immediate alternative to the expected advances of future fully quantum solutions, there are still scaling limitations. In parallel, Graph Neural Networks (GNN) have been recently adapted to solve combinatorial problems, showing competitive results and potentially high scalability due to their distributed nature. We propose a merging approach that aims at retaining both the accuracy exhibited by AMs and the representational flexibility and scalability of GNNs. Our model considers a compression step, followed by a supervised interaction where partial solutions obtained from the AM are used to guide local GNNs from where node feature representations are obtained and combined to initialize an additional GNN-based solver that handles the original graph's target problem. Intuitively, the AM can solve the combinatorial problem indirectly by infusing its knowledge into the GNN. Experiments on canonical optimization problems show that the idea is feasible, effectively allowing the AM to solve size problems beyond its original limits.
2501.05848
Isogeometric Analysis for 2D Magnetostatic Computations with Multi-level B\'{e}zier Extraction for Local Refinement
cs.CE
Local refinement is vital for efficient numerical simulations. In the context of Isogeometric Analysis (IGA), hierarchical B-splines have gained prominence. The work applies the methodology of truncated hierarchical B-splines (THB-splines) as they keep additional properties. The framework is further enriched with B\'{e}zier extraction, resulting in the multi-level B\'{e}zier extraction method. We apply this discretization method to 2D magnetostatic problems. The implementation is based on an open-source Octave/MATLAB IGA code called GeoPDEs, which allows us to compare our routines with globally refined spline models as well as locally refined ones where the solver does not rely on B\'{e}zier extraction.
2501.05851
Identity-aware Feature Decoupling Learning for Clothing-change Person Re-identification
cs.CV
Clothing-change person re-identification (CC Re-ID) has attracted increasing attention in recent years due to its application prospect. Most existing works struggle to adequately extract the ID-related information from the original RGB images. In this paper, we propose an Identity-aware Feature Decoupling (IFD) learning framework to mine identity-related features. Particularly, IFD exploits a dual stream architecture that consists of a main stream and an attention stream. The attention stream takes the clothing-masked images as inputs and derives the identity attention weights for effectively transferring the spatial knowledge to the main stream and highlighting the regions with abundant identity-related information. To eliminate the semantic gap between the inputs of two streams, we propose a clothing bias diminishing module specific to the main stream to regularize the features of clothing-relevant regions. Extensive experimental results demonstrate that our framework outperforms other baseline models on several widely-used CC Re-ID datasets.
2501.05852
MRI Patterns of the Hippocampus and Amygdala for Predicting Stages of Alzheimer's Progression: A Minimal Feature Machine Learning Framework
cs.CV cs.LG
Alzheimer's disease (AD) progresses through distinct stages, from early mild cognitive impairment (EMCI) to late mild cognitive impairment (LMCI) and eventually to AD. Accurate identification of these stages, especially distinguishing LMCI from EMCI, is crucial for developing pre-dementia treatments but remains challenging due to subtle and overlapping imaging features. This study proposes a minimal-feature machine learning framework that leverages structural MRI data, focusing on the hippocampus and amygdala as regions of interest. The framework addresses the curse of dimensionality through feature selection, utilizes region-specific voxel information, and implements innovative data organization to enhance classification performance by reducing noise. The methodology integrates dimensionality reduction techniques such as PCA and t-SNE with state-of-the-art classifiers, achieving the highest accuracy of 88.46%. This framework demonstrates the potential for efficient and accurate staging of AD progression while providing valuable insights for clinical applications.
2501.05855
ConSim: Measuring Concept-Based Explanations' Effectiveness with Automated Simulatability
cs.CL
Concept-based explanations work by mapping complex model computations to human-understandable concepts. Evaluating such explanations is very difficult, as it includes not only the quality of the induced space of possible concepts but also how effectively the chosen concepts are communicated to users. Existing evaluation metrics often focus solely on the former, neglecting the latter. We introduce an evaluation framework for measuring concept explanations via automated simulatability: a simulator's ability to predict the explained model's outputs based on the provided explanations. This approach accounts for both the concept space and its interpretation in an end-to-end evaluation. Human studies for simulatability are notoriously difficult to enact, particularly at the scale of a wide, comprehensive empirical evaluation (which is the subject of this work). We propose using large language models (LLMs) as simulators to approximate the evaluation and report various analyses to make such approximations reliable. Our method allows for scalable and consistent evaluation across various models and datasets. We report a comprehensive empirical evaluation using this framework and show that LLMs provide consistent rankings of explanation methods. Code available at https://github.com/AnonymousConSim/ConSim.
2501.05862
Language-Inspired Relation Transfer for Few-shot Class-Incremental Learning
cs.CV
Depicting novel classes with language descriptions by observing few-shot samples is inherent in human-learning systems. This lifelong learning capability helps to distinguish new knowledge from old ones through the increase of open-world learning, namely Few-Shot Class-Incremental Learning (FSCIL). Existing works to solve this problem mainly rely on the careful tuning of visual encoders, which shows an evident trade-off between the base knowledge and incremental ones. Motivated by human learning systems, we propose a new Language-inspired Relation Transfer (LRT) paradigm to understand objects by joint visual clues and text depictions, composed of two major steps. We first transfer the pretrained text knowledge to the visual domains by proposing a graph relation transformation module and then fuse the visual and language embedding by a text-vision prototypical fusion module. Second, to mitigate the domain gap caused by visual finetuning, we propose context prompt learning for fast domain alignment and imagined contrastive learning to alleviate the insufficient text data during alignment. With collaborative learning of domain alignments and text-image transfer, our proposed LRT outperforms the state-of-the-art models by over $13\%$ and $7\%$ on the final session of mini-ImageNet and CIFAR-100 FSCIL benchmarks.
2501.05867
Neural Network Verification is a Programming Language Challenge
cs.PL cs.LG cs.LO
Neural network verification is a new and rapidly developing field of research. So far, the main priority has been establishing efficient verification algorithms and tools, while proper support from the programming language perspective has been considered secondary or unimportant. Yet, there is mounting evidence that insights from the programming language community may make a difference in the future development of this domain. In this paper, we formulate neural network verification challenges as programming language challenges and suggest possible future solutions.
2501.05870
A Neighbor-based Approach to Pitch Ownership Models in Soccer
cs.LG
Pitch ownership models allow many types of analysis in soccer and provide valuable assistance to tactical analysts in understanding the game's dynamics. The novelty they provide over event-based analysis is that tracking data incorporates context that event-based data does not possess, like player positioning. This paper proposes a novel approach to building pitch ownership models in soccer games using the K-Nearest Neighbors (KNN) algorithm. Our approach provides a fast inference mechanism that can model different approaches to pitch control using the same algorithm. Despite its flexibility, it uses only three hyperparameters to tune the model, facilitating the tuning process for different player skill levels. The flexibility of the approach allows for the emulation of different methods available in the literature by adjusting a small number of parameters, including adjusting for different levels of uncertainty. In summary, the proposed model provides a new and more flexible strategy for building pitch ownership models, extending beyond just replicating existing algorithms, and can provide valuable insights for tactical analysts and open up new avenues for future research. We thoroughly visualize several examples demonstrating the presented models' strengths and weaknesses. The code is available at github.com/nvsclub/KNNPitchControl.
2501.05871
Collaborative Content Moderation in the Fediverse
cs.SI cs.LG cs.NI
The Fediverse, a group of interconnected servers providing a variety of interoperable services (e.g. micro-blogging in Mastodon) has gained rapid popularity. This sudden growth, partly driven by Elon Musk's acquisition of Twitter, has created challenges for administrators though. This paper focuses on one particular challenge: content moderation, e.g. the need to remove spam or hate speech. While centralized platforms like Facebook and Twitter rely on automated tools for moderation, their dependence on massive labeled datasets and specialized infrastructure renders them impractical for decentralized, low-resource settings like the Fediverse. In this work, we design and evaluate FedMod, a collaborative content moderation system based on federated learning. Our system enables servers to exchange parameters of partially trained local content moderation models with similar servers, creating a federated model shared among collaborating servers. FedMod demonstrates robust performance on three different content moderation tasks: harmful content detection, bot content detection, and content warning assignment, achieving average per-server macro-F1 scores of 0.71, 0.73, and 0.58, respectively.
2501.05874
VideoRAG: Retrieval-Augmented Generation over Video Corpus
cs.CV cs.AI cs.CL cs.IR cs.LG
Retrieval-Augmented Generation (RAG) is a powerful strategy to address the issue of generating factually incorrect outputs in foundation models by retrieving external knowledge relevant to queries and incorporating it into their generation process. However, existing RAG approaches have primarily focused on textual information, with some recent advancements beginning to consider images, and they largely overlook videos, a rich source of multimodal knowledge capable of representing events, processes, and contextual details more effectively than any other modality. While a few recent studies explore the integration of videos in the response generation process, they either predefine query-associated videos without retrieving them according to queries, or convert videos into the textual descriptions without harnessing their multimodal richness. To tackle these, we introduce VideoRAG, a novel framework that not only dynamically retrieves relevant videos based on their relevance with queries but also utilizes both visual and textual information of videos in the output generation. Further, to operationalize this, our method revolves around the recent advance of Large Video Language Models (LVLMs), which enable the direct processing of video content to represent it for retrieval and seamless integration of the retrieved videos jointly with queries. We experimentally validate the effectiveness of VideoRAG, showcasing that it is superior to relevant baselines.
2501.05880
TakuNet: an Energy-Efficient CNN for Real-Time Inference on Embedded UAV systems in Emergency Response Scenarios
cs.CV cs.PF
Designing efficient neural networks for embedded devices is a critical challenge, particularly in applications requiring real-time performance, such as aerial imaging with drones and UAVs for emergency responses. In this work, we introduce TakuNet, a novel light-weight architecture which employs techniques such as depth-wise convolutions and an early downsampling stem to reduce computational complexity while maintaining high accuracy. It leverages dense connections for fast convergence during training and uses 16-bit floating-point precision for optimization on embedded hardware accelerators. Experimental evaluation on two public datasets shows that TakuNet achieves near-state-of-the-art accuracy in classifying aerial images of emergency situations, despite its minimal parameter count. Real-world tests on embedded devices, namely Jetson Orin Nano and Raspberry Pi, confirm TakuNet's efficiency, achieving more than 650 fps on the 15W Jetson board, making it suitable for real-time AI processing on resource-constrained platforms and advancing the applicability of drones in emergency scenarios. The code and implementation details are publicly released.
2501.05882
Solving nonograms using Neural Networks
cs.AI cs.NE
Nonograms are logic puzzles in which cells in a grid must be colored or left blank according to the numbers that are located in its headers. In this study, we analyze different techniques to solve this type of logical problem using an Heuristic Algorithm, Genetic Algorithm, and Heuristic Algorithm with Neural Network. Furthermore, we generate a public dataset to train the neural networks. We published this dataset and the code of the algorithms. Combination of the heuristic algorithm with a neural network obtained the best results. From state of the art review, no previous works used neural network to solve nonograms, nor combined a network with other algorithms to accelerate the resolution process.
2501.05884
Text-to-Edit: Controllable End-to-End Video Ad Creation via Multimodal LLMs
cs.CV
The exponential growth of short-video content has ignited a surge in the necessity for efficient, automated solutions to video editing, with challenges arising from the need to understand videos and tailor the editing according to user requirements. Addressing this need, we propose an innovative end-to-end foundational framework, ultimately actualizing precise control over the final video content editing. Leveraging the flexibility and generalizability of Multimodal Large Language Models (MLLMs), we defined clear input-output mappings for efficient video creation. To bolster the model's capability in processing and comprehending video content, we introduce a strategic combination of a denser frame rate and a slow-fast processing technique, significantly enhancing the extraction and understanding of both temporal and spatial video information. Furthermore, we introduce a text-to-edit mechanism that allows users to achieve desired video outcomes through textual input, thereby enhancing the quality and controllability of the edited videos. Through comprehensive experimentation, our method has not only showcased significant effectiveness within advertising datasets, but also yields universally applicable conclusions on public datasets.
2501.05885
EDNet: Edge-Optimized Small Target Detection in UAV Imagery -- Faster Context Attention, Better Feature Fusion, and Hardware Acceleration
cs.CV cs.AI cs.LG
Detecting small targets in drone imagery is challenging due to low resolution, complex backgrounds, and dynamic scenes. We propose EDNet, a novel edge-target detection framework built on an enhanced YOLOv10 architecture, optimized for real-time applications without post-processing. EDNet incorporates an XSmall detection head and a Cross Concat strategy to improve feature fusion and multi-scale context awareness for detecting tiny targets in diverse environments. Our unique C2f-FCA block employs Faster Context Attention to enhance feature extraction while reducing computational complexity. The WIoU loss function is employed for improved bounding box regression. With seven model sizes ranging from Tiny to XL, EDNet accommodates various deployment environments, enabling local real-time inference and ensuring data privacy. Notably, EDNet achieves up to a 5.6% gain in mAP@50 with significantly fewer parameters. On an iPhone 12, EDNet variants operate at speeds ranging from 16 to 55 FPS, providing a scalable and efficient solution for edge-based object detection in challenging drone imagery. The source code and pre-trained models are available at: https://github.com/zsniko/EDNet.
2501.05891
Affordably Fine-tuned LLMs Provide Better Answers to Course-specific MCQs
cs.CL cs.AI
In education, the capability of generating human-like text of Large Language Models (LLMs) inspired work on how they can increase the efficiency of learning and teaching. We study the affordability of these models for educators and students by investigating how LLMs answer multiple-choice questions (MCQs) with respect to hardware constraints and refinement techniques. We explore this space by using generic pre-trained LLMs (the 7B, 13B, and 70B variants of LLaMA-2) to answer 162 undergraduate-level MCQs from a course on Programming Languages (PL) -- the MCQ dataset is a contribution of this work, which we make publicly available. Specifically, we dissect how different factors, such as using readily-available material -- (parts of) the course's textbook -- for fine-tuning and quantisation (to decrease resource usage) can change the accuracy of the responses. The main takeaway is that smaller textbook-based fine-tuned models outperform generic larger ones (whose pre-training requires conspicuous resources), making the usage of LLMs for answering MCQs resource- and material-wise affordable.
2501.05892
Beyond Flat Text: Dual Self-inherited Guidance for Visual Text Generation
cs.CV
In real-world images, slanted or curved texts, especially those on cans, banners, or badges, appear as frequently, if not more so, than flat texts due to artistic design or layout constraints. While high-quality visual text generation has become available with the advanced generative capabilities of diffusion models, these models often produce distorted text and inharmonious text background when given slanted or curved text layouts due to training data limitation. In this paper, we introduce a new training-free framework, STGen, which accurately generates visual texts in challenging scenarios (\eg, slanted or curved text layouts) while harmonizing them with the text background. Our framework decomposes the visual text generation process into two branches: (i) \textbf{Semantic Rectification Branch}, which leverages the ability in generating flat but accurate visual texts of the model to guide the generation of challenging scenarios. The generated latent of flat text is abundant in accurate semantic information related both to the text itself and its background. By incorporating this, we rectify the semantic information of the texts and harmonize the integration of the text with its background in complex layouts. (ii) \textbf{Structure Injection Branch}, which reinforces the visual text structure during inference. We incorporate the latent information of the glyph image, rich in glyph structure, as a new condition to further strengthen the text structure. To enhance image harmony, we also apply an effective combination method to merge the priors, providing a solid foundation for generation. Extensive experiments across a variety of visual text layouts demonstrate that our framework achieves superior accuracy and outstanding quality.
2501.05894
Text2Playlist: Generating Personalized Playlists from Text on Deezer
cs.IR cs.LG
The streaming service Deezer heavily relies on the search to help users navigate through its extensive music catalog. Nonetheless, it is primarily designed to find specific items and does not lead directly to a smooth listening experience. We present Text2Playlist, a stand-alone tool that addresses these limitations. Text2Playlist leverages generative AI, music information retrieval and recommendation systems to generate query-specific and personalized playlists, successfully deployed at scale.
2501.05901
Valley2: Exploring Multimodal Models with Scalable Vision-Language Design
cs.CV
Recently, vision-language models have made remarkable progress, demonstrating outstanding capabilities in various tasks such as image captioning and video understanding. We introduce Valley2, a novel multimodal large language model designed to enhance performance across all domains and extend the boundaries of practical applications in e-commerce and short video scenarios. Notably, Valley2 achieves state-of-the-art (SOTA) performance on e-commerce benchmarks, surpassing open-source models of similar size by a large margin (79.66 vs. 72.76). Additionally, Valley2 ranks second on the OpenCompass leaderboard among models with fewer than 10B parameters, with an impressive average score of 67.4. The code and model weights are open-sourced at https://github.com/bytedance/Valley.
2501.05903
Discovery of sustainable energy materials via the machine-learned material space
cond-mat.mtrl-sci cs.LG physics.comp-ph
Does a machine learning model actually gain an understanding of the material space? We answer this question in the affirmative on the example of the OptiMate model, a graph attention network trained to predict the optical properties of semiconductors and insulators. By applying the UMAP dimensionality reduction technique to its latent embeddings, we demonstrate that the model captures a nuanced and interpretable representation of the materials space, reflecting chemical and physical principles, without any user-induced bias. This enables clustering of almost 10,000 materials based on optical properties and chemical similarities. Beyond this understanding, we demonstrate how the learned material space can be used to identify more sustainable alternatives to critical materials in energy-related technologies, such as photovoltaics. These findings demonstrate the dual utility of machine learning models in materials science: Accurately predicting material properties while providing insights into the underlying materials space. The approach demonstrates the broader potential of leveraging learned materials spaces for the discovery and design of materials for diverse applications, and is easily applicable to any state-of-the-art machine learning model.
2501.05904
Binary Event-Driven Spiking Transformer
cs.CV
Transformer-based Spiking Neural Networks (SNNs) introduce a novel event-driven self-attention paradigm that combines the high performance of Transformers with the energy efficiency of SNNs. However, the larger model size and increased computational demands of the Transformer structure limit their practicality in resource-constrained scenarios. In this paper, we integrate binarization techniques into Transformer-based SNNs and propose the Binary Event-Driven Spiking Transformer, i.e. BESTformer. The proposed BESTformer can significantly reduce storage and computational demands by representing weights and attention maps with a mere 1-bit. However, BESTformer suffers from a severe performance drop from its full-precision counterpart due to the limited representation capability of binarization. To address this issue, we propose a Coupled Information Enhancement (CIE) method, which consists of a reversible framework and information enhancement distillation. By maximizing the mutual information between the binary model and its full-precision counterpart, the CIE method effectively mitigates the performance degradation of the BESTformer. Extensive experiments on static and neuromorphic datasets demonstrate that our method achieves superior performance to other binary SNNs, showcasing its potential as a compact yet high-performance model for resource-limited edge devices.
2501.05906
Q-MAML: Quantum Model-Agnostic Meta-Learning for Variational Quantum Algorithms
quant-ph cs.LG
In the Noisy Intermediate-Scale Quantum (NISQ) era, using variational quantum algorithms (VQAs) to solve optimization problems has become a key application. However, these algorithms face significant challenges, such as choosing an effective initial set of parameters and the limited quantum processing time that restricts the number of optimization iterations. In this study, we introduce a new framework for optimizing parameterized quantum circuits (PQCs) that employs a classical optimizer, inspired by Model-Agnostic Meta-Learning (MAML) technique. This approach aim to achieve better parameter initialization that ensures fast convergence. Our framework features a classical neural network, called Learner}, which interacts with a PQC using the output of Learner as an initial parameter. During the pre-training phase, Learner is trained with a meta-objective based on the quantum circuit cost function. In the adaptation phase, the framework requires only a few PQC updates to converge to a more accurate value, while the learner remains unchanged. This method is highly adaptable and is effectively extended to various Hamiltonian optimization problems. We validate our approach through experiments, including distribution function mapping and optimization of the Heisenberg XYZ Hamiltonian. The result implies that the Learner successfully estimates initial parameters that generalize across the problem space, enabling fast adaptation.
2501.05921
The New Anticipatory Governance Culture for Innovation: Regulatory Foresight, Regulatory Experimentation and Regulatory Learning
cs.CY cs.AI
With the rapid pace of technological innovation, traditional methods of policy formation and legislating are becoming conspicuously anachronistic. The need for regulatory choices to be made to counter the deadening effect of regulatory lag is more important to developing markets and fostering growth than achieving one off regulatory perfection. This article advances scholarship on innovation policy and the regulation of technological innovation in the European Union. It does so by considering what building an agile yet robust anticipatory governance regulatory culture involves. It systematically excavates a variety of tools and elements that are being put into use in inventive ways and argues that these need to be more cohesively and systemically integrated into the regulatory toolbox. Approaches covered include strategic foresight, the critical embrace of iterative policy development and regulatory learning in the face of uncertainty and the embrace of bottom up approaches to cocreation of policy such as Policy Labs and the testing and regulatory learning through pilot regulation and experimentation. The growing use of regulatory sandboxes as an EU policy tool to boost innovation and navigate regulatory complexity as seen in the EU AI Act is also probed
2501.05925
Navigating Tomorrow: Reliably Assessing Large Language Models Performance on Future Event Prediction
cs.CL cs.IR
Predicting future events is an important activity with applications across multiple fields and domains. For example, the capacity to foresee stock market trends, natural disasters, business developments, or political events can facilitate early preventive measures and uncover new opportunities. Multiple diverse computational methods for attempting future predictions, including predictive analysis, time series forecasting, and simulations have been proposed. This study evaluates the performance of several large language models (LLMs) in supporting future prediction tasks, an under-explored domain. We assess the models across three scenarios: Affirmative vs. Likelihood questioning, Reasoning, and Counterfactual analysis. For this, we create a dataset1 by finding and categorizing news articles based on entity type and its popularity. We gather news articles before and after the LLMs training cutoff date in order to thoroughly test and compare model performance. Our research highlights LLMs potential and limitations in predictive modeling, providing a foundation for future improvements.
2501.05926
LLMs Reproduce Stereotypes of Sexual and Gender Minorities
cs.CL
A large body of research has found substantial gender bias in NLP systems. Most of this research takes a binary, essentialist view of gender: limiting its variation to the categories _men_ and _women_, conflating gender with sex, and ignoring different sexual identities. But gender and sexuality exist on a spectrum, so in this paper we study the biases of large language models (LLMs) towards sexual and gender minorities beyond binary categories. Grounding our study in a widely used psychological framework -- the Stereotype Content Model -- we demonstrate that English-language survey questions about social perceptions elicit more negative stereotypes of sexual and gender minorities from LLMs, just as they do from humans. We then extend this framework to a more realistic use case: text generation. Our analysis shows that LLMs generate stereotyped representations of sexual and gender minorities in this setting, raising concerns about their capacity to amplify representational harms in creative writing, a widely promoted use case.
2501.05927
Expressing One's Identity Online: Left-Right and cross EU-country variation in self-representation in social media
cs.SI
We examine how social media users from eight European Union (EU) member states express their socio-political identities, focusing on users' online self-presentation and group identity cues conveyed through bios. Our goal is to explore commonalities and differences in topics discussed in social media profiles, across Left-and Right-wing user groups, within and across EU countries. Through a novel approach we map how identity-related discourse varies by country and political orientation, revealing how group identity is expressed within the EU. We find that topics related to democracy, national way of life, and decentralization emerge as particularly divisive, showing considerable variation both within and between EU countries. A subset of topics, which includes education, environmentalism, sustainability, equality, freedom & human rights, and traditional morality, among others, clearly differentiate Left-from Right-leaning user groups. These partisan topics are relevant as they could be leveraged for mobilizing ideological groups and highlight Left-Right identitarian differences at the EU level. Finally, we show that our Left-Right identity similarity metrics reflect aspects of real-world political fragmentation, which are closely aligned to the perceptions of political conflict intensity by country, as measured by the 2022 PEW survey.
2501.05928
Towards Backdoor Stealthiness in Model Parameter Space
cs.CR cs.AI
Recent research on backdoor stealthiness focuses mainly on indistinguishable triggers in input space and inseparable backdoor representations in feature space, aiming to circumvent backdoor defenses that examine these respective spaces. However, existing backdoor attacks are typically designed to resist a specific type of backdoor defense without considering the diverse range of defense mechanisms. Based on this observation, we pose a natural question: Are current backdoor attacks truly a real-world threat when facing diverse practical defenses? To answer this question, we examine 12 common backdoor attacks that focus on input-space or feature-space stealthiness and 17 diverse representative defenses. Surprisingly, we reveal a critical blind spot: Backdoor attacks designed to be stealthy in input and feature spaces can be mitigated by examining backdoored models in parameter space. To investigate the underlying causes behind this common vulnerability, we study the characteristics of backdoor attacks in the parameter space. Notably, we find that input- and feature-space attacks introduce prominent backdoor-related neurons in parameter space, which are not thoroughly considered by current backdoor attacks. Taking comprehensive stealthiness into account, we propose a novel supply-chain attack called Grond. Grond limits the parameter changes by a simple yet effective module, Adversarial Backdoor Injection (ABI), which adaptively increases the parameter-space stealthiness during the backdoor injection. Extensive experiments demonstrate that Grond outperforms all 12 backdoor attacks against state-of-the-art (including adaptive) defenses on CIFAR-10, GTSRB, and a subset of ImageNet. In addition, we show that ABI consistently improves the effectiveness of common backdoor attacks.
2501.05931
Environment Modeling for Service Robots From a Task Execution Perspective
cs.RO
Service robots are increasingly entering the home to provide domestic tasks for residents. However, when working in an open, dynamic, and unstructured home environment, service robots still face challenges such as low intelligence for task execution and poor long-term autonomy (LTA), which has limited their deployment. As the basis of robotic task execution, environment modeling has attracted significant attention. This integrates core technologies such as environment perception, understanding, and representation to accurately recognize environmental information. This paper presents a comprehensive survey of environmental modeling from a new task-executionoriented perspective. In particular, guided by the requirements of robots in performing domestic service tasks in the home environment, we systematically review the progress that has been made in task-execution-oriented environmental modeling in four respects: 1) localization, 2) navigation, 3) manipulation, and 4) LTA. Current challenges are discussed, and potential research opportunities are also highlighted.
2501.05932
DiffuSETS: 12-lead ECG Generation Conditioned on Clinical Text Reports and Patient-Specific Information
cs.LG cs.AI
Heart disease remains a significant threat to human health. As a non-invasive diagnostic tool, the electrocardiogram (ECG) is one of the most widely used methods for cardiac screening. However, the scarcity of high-quality ECG data, driven by privacy concerns and limited medical resources, creates a pressing need for effective ECG signal generation. Existing approaches for generating ECG signals typically rely on small training datasets, lack comprehensive evaluation frameworks, and overlook potential applications beyond data augmentation. To address these challenges, we propose DiffuSETS, a novel framework capable of generating ECG signals with high semantic alignment and fidelity. DiffuSETS accepts various modalities of clinical text reports and patient-specific information as inputs, enabling the creation of clinically meaningful ECG signals. Additionally, to address the lack of standardized evaluation in ECG generation, we introduce a comprehensive benchmarking methodology to assess the effectiveness of generative models in this domain. Our model achieve excellent results in tests, proving its superiority in the task of ECG generation. Furthermore, we showcase its potential to mitigate data scarcity while exploring novel applications in cardiology education and medical knowledge discovery, highlighting the broader impact of our work.
2501.05933
Weakly Supervised Segmentation of Hyper-Reflective Foci with Compact Convolutional Transformers and SAM2
cs.CV
Weakly supervised segmentation has the potential to greatly reduce the annotation effort for training segmentation models for small structures such as hyper-reflective foci (HRF) in optical coherence tomography (OCT). However, most weakly supervised methods either involve a strong downsampling of input images, or only achieve localization at a coarse resolution, both of which are unsatisfactory for small structures. We propose a novel framework that increases the spatial resolution of a traditional attention-based Multiple Instance Learning (MIL) approach by using Layer-wise Relevance Propagation (LRP) to prompt the Segment Anything Model (SAM~2), and increases recall with iterative inference. Moreover, we demonstrate that replacing MIL with a Compact Convolutional Transformer (CCT), which adds a positional encoding, and permits an exchange of information between different regions of the OCT image, leads to a further and substantial increase in segmentation accuracy.
2501.05934
Encoded Spatial Attribute in Multi-Tier Federated Learning
cs.LG cs.DC
This research presents an Encoded Spatial Multi-Tier Federated Learning approach for a comprehensive evaluation of aggregated models for geospatial data. In the client tier, encoding spatial information is introduced to better predict the target outcome. The research aims to assess the performance of these models across diverse datasets and spatial attributes, highlighting variations in predictive accuracy. Using evaluation metrics such as accuracy, our research reveals insights into the complexities of spatial granularity and the challenges of capturing underlying patterns in the data. We extended the scope of federated learning (FL) by having multi-tier along with the functionality of encoding spatial attributes. Our N-tier FL approach used encoded spatial data to aggregate in different tiers. We obtained multiple models that predicted the different granularities of spatial data. Our findings underscore the need for further research to improve predictive accuracy and model generalization, with potential avenues including incorporating additional features, refining model architectures, and exploring alternative modeling approaches. Our experiments have several tiers representing different levels of spatial aspects. We obtained accuracy of 75.62% and 89.52% for the global model without having to train the model using the data constituted with the designated tier. The research also highlights the importance of the proposed approach in real-time applications.
2501.05936
A Multimodal Dataset for Enhancing Industrial Task Monitoring and Engagement Prediction
cs.CV
Detecting and interpreting operator actions, engagement, and object interactions in dynamic industrial workflows remains a significant challenge in human-robot collaboration research, especially within complex, real-world environments. Traditional unimodal methods often fall short of capturing the intricacies of these unstructured industrial settings. To address this gap, we present a novel Multimodal Industrial Activity Monitoring (MIAM) dataset that captures realistic assembly and disassembly tasks, facilitating the evaluation of key meta-tasks such as action localization, object interaction, and engagement prediction. The dataset comprises multi-view RGB, depth, and Inertial Measurement Unit (IMU) data collected from 22 sessions, amounting to 290 minutes of untrimmed video, annotated in detail for task performance and operator behavior. Its distinctiveness lies in the integration of multiple data modalities and its emphasis on real-world, untrimmed industrial workflows-key for advancing research in human-robot collaboration and operator monitoring. Additionally, we propose a multimodal network that fuses RGB frames, IMU data, and skeleton sequences to predict engagement levels during industrial tasks. Our approach improves the accuracy of recognizing engagement states, providing a robust solution for monitoring operator performance in dynamic industrial environments. The dataset and code can be accessed from https://github.com/navalkishoremehta95/MIAM/.
2501.05942
Soft regression trees: a model variant and a decomposition training algorithm
cs.LG math.OC
Decision trees are widely used for classification and regression tasks in a variety of application fields due to their interpretability and good accuracy. During the past decade, growing attention has been devoted to globally optimized decision trees with deterministic or soft splitting rules at branch nodes, which are trained by optimizing the error function over all the tree parameters. In this work, we propose a new variant of soft multivariate regression trees (SRTs) where, for every input vector, the prediction is defined as the linear regression associated to a single leaf node, namely, the leaf node obtained by routing the input vector from the root along the branches with higher probability. SRTs exhibit the conditional computational property, i.e., each prediction depends on a small number of nodes (parameters), and our nonlinear optimization formulation for training them is amenable to decomposition. After showing a universal approximation result for SRTs, we present a decomposition training algorithm including a clustering-based initialization procedure and a heuristic for reassigning the input vectors along the tree. Under mild assumptions, we establish asymptotic convergence guarantees. Experiments on 15 wellknown datasets indicate that our SRTs and decomposition algorithm yield higher accuracy and robustness compared with traditional soft regression trees trained using the nonlinear optimization formulation of Blanquero et al., and a significant reduction in training times as well as a slightly better average accuracy compared with the mixed-integer optimization approach of Bertsimas and Dunn. We also report a comparison with the Random Forest ensemble method.
2501.05943
Koopman-Based Model Predictive Control of Functional Electrical Stimulation for Ankle Dorsiflexion and Plantarflexion Assistance
eess.SY cs.SY
Functional Electrical Stimulation (FES) can be an effective tool to augment paretic muscle function and restore normal ankle function. Our approach incorporates a real-time, data-driven Model Predictive Control (MPC) scheme, built upon a Koopman operator theory (KOT) framework. This framework adeptly captures the complex nonlinear dynamics of ankle motion in a linearized form, enabling application of linear control approaches for highly nonlinear FES-actuated dynamics. Utilizing inertial measurement units (IMUs), our method accurately predicts the FES-induced ankle movements, while accounting for nonlinear muscle actuation dynamics, including the muscle activation for both plantarflexors, and dorsiflexors (Tibialis Anterior (TA)). The linear prediction model derived through KOT allowed us to formulate the MPC problem with linear state space dynamics, enhancing the real-time feasibility, precision and adaptability of the FES driven control. The effectiveness and applicability of our approach have been demonstrated through comprehensive simulations and experimental trials, including three participants with no disability and a participant with Multiple Sclerosis. Our findings highlight the potential of a KOT-based MPC approach for FES based gait assistance that offers effective and personalized assistance for individuals with gait impairment conditions.
2501.05945
Reusable specimen-level inference in computational pathology
eess.IV cs.CV q-bio.TO
Foundation models for computational pathology have shown great promise for specimen-level tasks and are increasingly accessible to researchers. However, specimen-level models built on these foundation models remain largely unavailable, hindering their broader utility and impact. To address this gap, we developed SpinPath, a toolkit designed to democratize specimen-level deep learning by providing a zoo of pretrained specimen-level models, a Python-based inference engine, and a JavaScript-based inference platform. We demonstrate the utility of SpinPath in metastasis detection tasks across nine foundation models. SpinPath may foster reproducibility, simplify experimentation, and accelerate the adoption of specimen-level deep learning in computational pathology research.
2501.05946
Coverage and Spectral Efficiency of NOMA-Enabled LEO Satellite Networks with Ordering Schemes
eess.SP cs.IT cs.SY eess.SY math.IT
This paper investigates an analytical model for low-earth orbit (LEO) multi-satellite downlink non-orthogonal multiple access (NOMA) networks. The satellites transmit data to multiple NOMA user terminals (UTs), each employing successive interference cancellation (SIC) for decoding. Two ordering schemes are adopted for NOMA-enabled LEO satellite networks, i.e., mean signal power (MSP)-based ordering and instantaneous-signal-to-inter-satellite-interference-plus-noise ratio (ISINR)-based ordering. For each ordering scheme, we derive the coverage probabilities of UTs under different channel conditions. Moreover, we discuss how coverage is influenced by SIC, main-lobe gain, and tradeoffs between the number of satellites and their altitudes. Additionally, two user fairness-based power allocation (PA) schemes are considered, and PA coefficients with the optimal number of UTs that maximize their sum spectral efficiency (SE) are studied. Simulation results show that there exists a maximum signal-to-inter-satellite-interference-plus-noise ratio (SINR) threshold for each PA scheme that ensures the operation of NOMA in LEO satellite networks, and the benefit of NOMA only exists when the target SINR is below a certain threshold. Compared with orthogonal multiple access (OMA), NOMA increases UTs' sum SE by as much as 35\%. Furthermore, for most SINR thresholds, the sum SE increases with the number of UTs to the highest value, whilst the maximum sum SE is obtained when there are two UTs.
2501.05948
Universal-2-TF: Robust All-Neural Text Formatting for ASR
cs.CL
This paper introduces an all-neural text formatting (TF) model designed for commercial automatic speech recognition (ASR) systems, encompassing punctuation restoration (PR), truecasing, and inverse text normalization (ITN). Unlike traditional rule-based or hybrid approaches, this method leverages a two-stage neural architecture comprising a multi-objective token classifier and a sequence-to-sequence (seq2seq) model. This design minimizes computational costs and reduces hallucinations while ensuring flexibility and robustness across diverse linguistic entities and text domains. Developed as part of the Universal-2 ASR system, the proposed method demonstrates superior performance in TF accuracy, computational efficiency, and perceptual quality, as validated through comprehensive evaluations using both objective and subjective methods. This work underscores the importance of holistic TF models in enhancing ASR usability in practical settings.
2501.05952
Scalable Vision Language Model Training via High Quality Data Curation
cs.CV cs.CL
In this paper, we introduce SAIL-VL (ScAlable Vision Language Model TraIning via High QuaLity Data Curation), an open-source vision language model (VLM) series achieving state-of-the-art (SOTA) performance in 2B and 8B parameters. The following three key improvements contribute to SAIL-VL's leading performance: (1) Scalable high-quality visual understanding data construction: We implement a data construction pipeline to enable hundred-million-scale high-quality recaption data annotation, and the resulted dataset SAIL-Caption is validated to be of the highest data quality compared with opensource alternatives. (2) Scalable Pretraining with High-Quality Visual Understanding Data: We scale SAIL-VL's pretraining budget up to 655B tokens and show that even a 2B VLM benefits from scaled up training data sizes, exhibiting expected data size scaling laws in visual understanding and instruction following performance. (3) Scalable SFT via data quantity and complexity scaling: We curate a high-quality SFT dataset collection which outperforms opensource alternatives in data quantity scaling effectiveness. We also demonstrate that training with progressively higher-complexity data surpasses baseline one-stage training by a large margin. SAIL-VL series models achieve the highest average score in 18 widely used VLM benchmarks in our evaluation, with the 2B model takes the top position over VLMs of comparable sizes on OpenCompass 2024 (https://rank.opencompass.org.cn/leaderboard-multimodal) demonstrating robust visual comprehension abilities. SAIL-VL series models are released at HuggingFace (https://huggingface.co/BytedanceDouyinContent).
2501.05961
Swin-X2S: Reconstructing 3D Shape from 2D Biplanar X-ray with Swin Transformers
cs.CV eess.IV
The conversion from 2D X-ray to 3D shape holds significant potential for improving diagnostic efficiency and safety. However, existing reconstruction methods often rely on hand-crafted features, manual intervention, and prior knowledge, resulting in unstable shape errors and additional processing costs. In this paper, we introduce Swin-X2S, an end-to-end deep learning method for directly reconstructing 3D segmentation and labeling from 2D biplanar orthogonal X-ray images. Swin-X2S employs an encoder-decoder architecture: the encoder leverages 2D Swin Transformer for X-ray information extraction, while the decoder employs 3D convolution with cross-attention to integrate structural features from orthogonal views. A dimension-expanding module is introduced to bridge the encoder and decoder, ensuring a smooth conversion from 2D pixels to 3D voxels. We evaluate proposed method through extensive qualitative and quantitative experiments across nine publicly available datasets covering four anatomies (femur, hip, spine, and rib), with a total of 54 categories. Significant improvements over previous methods have been observed not only in the segmentation and labeling metrics but also in the clinically relevant parameters that are of primary concern in practical applications, which demonstrates the promise of Swin-X2S to provide an effective option for anatomical shape reconstruction in clinical scenarios. Code implementation is available at: \url{https://github.com/liukuan5625/Swin-X2S}.
2501.05962
Effective faking of verbal deception detection with target-aligned adversarial attacks
cs.CL cs.AI
Background: Deception detection through analysing language is a promising avenue using both human judgments and automated machine learning judgments. For both forms of credibility assessment, automated adversarial attacks that rewrite deceptive statements to appear truthful pose a serious threat. Methods: We used a dataset of 243 truthful and 262 fabricated autobiographical stories in a deception detection task for humans and machine learning models. A large language model was tasked to rewrite deceptive statements so that they appear truthful. In Study 1, humans who made a deception judgment or used the detailedness heuristic and two machine learning models (a fine-tuned language model and a simple n-gram model) judged original or adversarial modifications of deceptive statements. In Study 2, we manipulated the target alignment of the modifications, i.e. tailoring the attack to whether the statements would be assessed by humans or computer models. Results: When adversarial modifications were aligned with their target, human (d=-0.07 and d=-0.04) and machine judgments (51% accuracy) dropped to the chance level. When the attack was not aligned with the target, both human heuristics judgments (d=0.30 and d=0.36) and machine learning predictions (63-78%) were significantly better than chance. Conclusions: Easily accessible language models can effectively help anyone fake deception detection efforts both by humans and machine learning models. Robustness against adversarial modifications for humans and machines depends on that target alignment. We close with suggestions on advancing deception research with adversarial attack designs.
2501.05963
Finnish SQuAD: A Simple Approach to Machine Translation of Span Annotations
cs.CL
We apply a simple method to machine translate datasets with span-level annotation using the DeepL MT service and its ability to translate formatted documents. Using this method, we produce a Finnish version of the SQuAD2.0 question answering dataset and train QA retriever models on this new dataset. We evaluate the quality of the dataset and more generally the MT method through direct evaluation, indirect comparison to other similar datasets, a backtranslation experiment, as well as through the performance of downstream trained QA models. In all these evaluations, we find that the method of transfer is not only simple to use but produces consistently better translated data. Given its good performance on the SQuAD dataset, it is likely the method can be used to translate other similar span-annotated datasets for other tasks and languages as well. All code and data is available under an open license: data at HuggingFace TurkuNLP/squad_v2_fi, code on GitHub TurkuNLP/squad2-fi, and model at HuggingFace TurkuNLP/bert-base-finnish-cased-squad2.
2501.05964
Recommender Systems for Social Good: The Role of Accountability and Sustainability
cs.IR
This work examines the role of recommender systems in promoting sustainability, social responsibility, and accountability, with a focus on alignment with the United Nations Sustainable Development Goals (SDGs). As recommender systems become increasingly integrated into daily interactions, they must go beyond personalization to support responsible consumption, reduce environmental impact, and foster social good. We explore strategies to mitigate the carbon footprint of recommendation models, ensure fairness, and implement accountability mechanisms. By adopting these approaches, recommender systems can contribute to sustainable and socially beneficial outcomes, aligning technological advancements with the SDGs focused on environmental sustainability and social well-being.
2501.05965
Model Inversion in Split Learning for Personalized LLMs: New Insights from Information Bottleneck Theory
cs.LG
Personalized Large Language Models (LLMs) have become increasingly prevalent, showcasing the impressive capabilities of models like GPT-4. This trend has also catalyzed extensive research on deploying LLMs on mobile devices. Feasible approaches for such edge-cloud deployment include using split learning. However, previous research has largely overlooked the privacy leakage associated with intermediate representations transmitted from devices to servers. This work is the first to identify model inversion attacks in the split learning framework for LLMs, emphasizing the necessity of secure defense. For the first time, we introduce mutual information entropy to understand the information propagation of Transformer-based LLMs and assess privacy attack performance for LLM blocks. To address the issue of representations being sparser and containing less information than embeddings, we propose a two-stage attack system in which the first part projects representations into the embedding space, and the second part uses a generative model to recover text from these embeddings. This design breaks down the complexity and achieves attack scores of 38%-75% in various scenarios, with an over 60% improvement over the SOTA. This work comprehensively highlights the potential privacy risks during the deployment of personalized LLMs on the edge side.
2501.05966
Towards Early Prediction of Self-Supervised Speech Model Performance
cs.SD cs.CL cs.LG eess.AS
In Self-Supervised Learning (SSL), pre-training and evaluation are resource intensive. In the speech domain, current indicators of the quality of SSL models during pre-training, such as the loss, do not correlate well with downstream performance. Consequently, it is often difficult to gauge the final downstream performance in a cost efficient manner during pre-training. In this work, we propose unsupervised efficient methods that give insights into the quality of the pre-training of SSL speech models, namely, measuring the cluster quality and rank of the embeddings of the SSL model. Results show that measures of cluster quality and rank correlate better with downstream performance than the pre-training loss with only one hour of unlabeled audio, reducing the need for GPU hours and labeled data in SSL model evaluation.
2501.05970
A Brain Age Residual Biomarker (BARB): Leveraging MRI-Based Models to Detect Latent Health Conditions in U.S. Veterans
cs.LG
Age prediction using brain imaging, such as MRIs, has achieved promising results, with several studies identifying the model's residual as a potential biomarker for chronic disease states. In this study, we developed a brain age predictive model using a dataset of 1,220 U.S. veterans (18--80 years) and convolutional neural networks (CNNs) trained on two-dimensional slices of axial T2-weighted fast spin-echo and T2-weighted fluid attenuated inversion recovery MRI images. The model, incorporating a degree-3 polynomial ensemble, achieved an $R^{2}$ of 0.816 on the testing set. Images were acquired at the level of the anterior commissure and the frontal horns of the lateral ventricles. Residual analysis was performed to assess its potential as a biomarker for five ICD-coded conditions: hypertension (HTN), diabetes mellitus (DM), mild traumatic brain injury (mTBI), illicit substance abuse/dependence (SAD), and alcohol abuse/dependence (AAD). Residuals grouped by the number of ICD-coded conditions demonstrated different trends that were statistically significant ($p = 0.002$), suggesting a relationship between disease states and predicted brain age. This association was particularly pronounced in patients over 49 years, where negative residuals (indicating advanced brain aging) correlated with the presence of multiple ICD codes. These findings support the potential of residuals as biomarkers for detecting latent health conditions.
2501.05981
Hermit Kingdom Through the Lens of Multiple Perspectives: A Case Study of LLM Hallucination on North Korea
cs.CL
Hallucination in large language models (LLMs) remains a significant challenge for their safe deployment, particularly due to its potential to spread misinformation. Most existing solutions address this challenge by focusing on aligning the models with credible sources or by improving how models communicate their confidence (or lack thereof) in their outputs. While these measures may be effective in most contexts, they may fall short in scenarios requiring more nuanced approaches, especially in situations where access to accurate data is limited or determining credible sources is challenging. In this study, we take North Korea - a country characterised by an extreme lack of reliable sources and the prevalence of sensationalist falsehoods - as a case study. We explore and evaluate how some of the best-performing multilingual LLMs and specific language-based models generate information about North Korea in three languages spoken in countries with significant geo-political interests: English (United States, United Kingdom), Korean (South Korea), and Mandarin Chinese (China). Our findings reveal significant differences, suggesting that the choice of model and language can lead to vastly different understandings of North Korea, which has important implications given the global security challenges the country poses.
2501.05982
Deep Variational Sequential Monte Carlo for High-Dimensional Observations
cs.LG eess.SP
Sequential Monte Carlo (SMC), or particle filtering, is widely used in nonlinear state-space systems, but its performance often suffers from poorly approximated proposal and state-transition distributions. This work introduces a differentiable particle filter that leverages the unsupervised variational SMC objective to parameterize the proposal and transition distributions with a neural network, designed to learn from high-dimensional observations. Experimental results demonstrate that our approach outperforms established baselines in tracking the challenging Lorenz attractor from high-dimensional and partial observations. Furthermore, an evidence lower bound based evaluation indicates that our method offers a more accurate representation of the posterior distribution.
2501.05984
The Safe Trusted Autonomy for Responsible Space Program
eess.SY cs.SY
The Safe Trusted Autonomy for Responsible Space (STARS) program aims to advance autonomy technologies for space by leveraging machine learning technologies while mitigating barriers to trust, such as uncertainty, opaqueness, brittleness, and inflexibility. This paper presents the achievements and lessons learned from the STARS program in integrating reinforcement learning-based multi-satellite control, run time assurance approaches, and flexible human-autonomy teaming interfaces, into a new integrated testing environment for collaborative autonomous satellite systems. The primary results describe analysis of the reinforcement learning multi-satellite control and run time assurance algorithms. These algorithms are integrated into a prototype human-autonomy interface using best practices from human-autonomy trust literature, however detailed analysis of the effectiveness is left to future work. References are provided with additional detailed results of individual experiments.
2501.05987
Comparing Self-Supervised Learning Models Pre-Trained on Human Speech and Animal Vocalizations for Bioacoustics Processing
cs.LG eess.AS
Self-supervised learning (SSL) foundation models have emerged as powerful, domain-agnostic, general-purpose feature extractors applicable to a wide range of tasks. Such models pre-trained on human speech have demonstrated high transferability for bioacoustic processing. This paper investigates (i) whether SSL models pre-trained directly on animal vocalizations offer a significant advantage over those pre-trained on speech, and (ii) whether fine-tuning speech-pretrained models on automatic speech recognition (ASR) tasks can enhance bioacoustic classification. We conduct a comparative analysis using three diverse bioacoustic datasets and two different bioacoustic tasks. Results indicate that pre-training on bioacoustic data provides only marginal improvements over speech-pretrained models, with comparable performance in most scenarios. Fine-tuning on ASR tasks yields mixed outcomes, suggesting that the general-purpose representations learned during SSL pre-training are already well-suited for bioacoustic tasks. These findings highlight the robustness of speech-pretrained SSL models for bioacoustics and imply that extensive fine-tuning may not be necessary for optimal performance.
2501.05989
Addressing speaker gender bias in large scale speech translation systems
cs.CL cs.AI
This study addresses the issue of speaker gender bias in Speech Translation (ST) systems, which can lead to offensive and inaccurate translations. The masculine bias often found in large-scale ST systems is typically perpetuated through training data derived from Machine Translation (MT) systems. Our approach involves two key steps. First, we employ Large Language Models (LLMs) to rectify translations based on the speaker's gender in a cost-effective manner. Second, we fine-tune the ST model with the corrected data, enabling the model to generate gender-specific translations directly from audio cues, without the need for explicit gender input. Additionally, we propose a three-mode fine-tuned model for scenarios where the speaker's gender is either predefined or should not be inferred from speech cues. We demonstrate a 70% improvement in translations for female speakers compared to our baseline and other large-scale ST systems, such as Seamless M4T and Canary, on the MuST-SHE test set.
2501.05990
Constraining constructions with WordNet: pros and cons for the semantic annotation of fillers in the Italian Constructicon
cs.CL
The paper discusses the role of WordNet-based semantic classification in the formalization of constructions, and more specifically in the semantic annotation of schematic fillers, in the Italian Constructicon. We outline how the Italian Constructicon project uses Open Multilingual WordNet topics to represent semantic features and constraints of constructions.
2501.05991
An Attention-Guided Deep Learning Approach for Classifying 39 Skin Lesion Types
eess.IV cs.CV cs.LG
The skin, as the largest organ of the human body, is vulnerable to a diverse array of conditions collectively known as skin lesions, which encompass various dermatoses. Diagnosing these lesions presents significant challenges for medical practitioners due to the subtle visual differences that are often imperceptible to the naked eye. While not all skin lesions are life-threatening, certain types can act as early indicators of severe diseases, including skin cancers, underscoring the critical need for timely and accurate diagnostic methods. Deep learning algorithms have demonstrated remarkable potential in facilitating the early detection and prognosis of skin lesions. This study advances the field by curating a comprehensive and diverse dataset comprising 39 categories of skin lesions, synthesized from five publicly available datasets. Using this dataset, the performance of five state-of-the-art deep learning models -- MobileNetV2, Xception, InceptionV3, EfficientNetB1, and Vision Transformer - is rigorously evaluated. To enhance the accuracy and robustness of these models, attention mechanisms such as the Efficient Channel Attention (ECA) and the Convolutional Block Attention Module (CBAM) are incorporated into their architectures. Comprehensive evaluation across multiple performance metrics reveals that the Vision Transformer model integrated with CBAM outperforms others, achieving an accuracy of 93.46%, precision of 94%, recall of 93%, F1-score of 93%, and specificity of 93.67%. These results underscore the significant potential of the proposed system in supporting medical professionals with accurate and efficient prognostic tools for diagnosing a broad spectrum of skin lesions. The dataset and code used in this study can be found at https://github.com/akabircs/Skin-Lesions-Classification.
2501.05994
On the Interaction in Transient Stability of Two-Inverter Power Systems containing GFL inverter Using Manifold Method
eess.SY cs.SY
Many renewable energy resources are integrated into power systems via grid-following (GFL) inverters which rely on a phase-locked loop (PLL) for grid synchronization. During severe grid faults, GFL inverters are vulnerable to transient instability, often leading to disconnection from the grid. This paper aims to elucidate the interaction mechanisms and define the stability boundaries of systems of two inverters, including GFL, grid-forming (GFM), or grid-supporting (GSP) inverters. First, the generalized large-signal expression for the two-inverter system under various inverter combinations is derived, revealing that no energy function exists for systems containing GFL inverters. This implies that the traditional direct method cannot be applied to such systems. To overcome these challenges, a manifold method is employed to precisely determine the domain of attraction (DOA) of the system, and the transient stability margin is assessed by a new metric termed the critical clearing radius (CCR). A case study of the two-inverter system under various inverter combinations is conducted to explore large-signal interactions across different scenarios. Manifold analysis and simulation results reveal that GSP inverters using PLL for grid synchronization exhibit behavior similar to GFM inverters when the droop coefficients in the terminal voltage control loop (TVC) are sufficiently large. Compared to GFL inverters, GSP inverters incorporating a TVC significantly enhances the transient stability of other inverters. In the STATCOM case, the optimal placement of the STATCOM, realized by GSP or GFM inverters, is identified to be at the midpoint of a transmission line. All findings in this paper are validated through electromagnetic transient (EMT) simulations
2501.05997
Minimizing Occlusion Effect on Multi-View Camera Perception in BEV with Multi-Sensor Fusion
cs.CV
Autonomous driving technology is rapidly evolving, offering the potential for safer and more efficient transportation. However, the performance of these systems can be significantly compromised by the occlusion on sensors due to environmental factors like dirt, dust, rain, and fog. These occlusions severely affect vision-based tasks such as object detection, vehicle segmentation, and lane recognition. In this paper, we investigate the impact of various kinds of occlusions on camera sensor by projecting their effects from multi-view camera images of the nuScenes dataset into the Bird's-Eye View (BEV) domain. This approach allows us to analyze how occlusions spatially distribute and influence vehicle segmentation accuracy within the BEV domain. Despite significant advances in sensor technology and multi-sensor fusion, a gap remains in the existing literature regarding the specific effects of camera occlusions on BEV-based perception systems. To address this gap, we use a multi-sensor fusion technique that integrates LiDAR and radar sensor data to mitigate the performance degradation caused by occluded cameras. Our findings demonstrate that this approach significantly enhances the accuracy and robustness of vehicle segmentation tasks, leading to more reliable autonomous driving systems.
2501.06000
Self-Supervised Partial Cycle-Consistency for Multi-View Matching
cs.CV
Matching objects across partially overlapping camera views is crucial in multi-camera systems and requires a view-invariant feature extraction network. Training such a network with cycle-consistency circumvents the need for labor-intensive labeling. In this paper, we extend the mathematical formulation of cycle-consistency to handle partial overlap. We then introduce a pseudo-mask which directs the training loss to take partial overlap into account. We additionally present several new cycle variants that complement each other and present a time-divergent scene sampling scheme that improves the data input for this self-supervised setting. Cross-camera matching experiments on the challenging DIVOTrack dataset show the merits of our approach. Compared to the self-supervised state-of-the-art, we achieve a 4.3 percentage point higher F1 score with our combined contributions. Our improvements are robust to reduced overlap in the training data, with substantial improvements in challenging scenes that need to make few matches between many people. Self-supervised feature networks trained with our method are effective at matching objects in a range of multi-camera settings, providing opportunities for complex tasks like large-scale multi-camera scene understanding.
2501.06002
DeltaGNN: Graph Neural Network with Information Flow Control
cs.LG
Graph Neural Networks (GNNs) are popular deep learning models designed to process graph-structured data through recursive neighborhood aggregations in the message passing process. When applied to semi-supervised node classification, the message-passing enables GNNs to understand short-range spatial interactions, but also causes them to suffer from over-smoothing and over-squashing. These challenges hinder model expressiveness and prevent the use of deeper models to capture long-range node interactions (LRIs) within the graph. Popular solutions for LRIs detection are either too expensive to process large graphs due to high time complexity or fail to generalize across diverse graph structures. To address these limitations, we propose a mechanism called \emph{information flow control}, which leverages a novel connectivity measure, called \emph{information flow score}, to address over-smoothing and over-squashing with linear computational overhead, supported by theoretical evidence. Finally, to prove the efficacy of our methodology we design DeltaGNN, the first scalable and generalizable approach for detecting long-range and short-range interactions. We benchmark our model across 10 real-world datasets, including graphs with varying sizes, topologies, densities, and homophilic ratios, showing superior performance with limited computational complexity. The implementation of the proposed methods are publicly available at https://github.com/basiralab/DeltaGNN.