id
stringlengths
9
16
title
stringlengths
4
278
categories
stringlengths
5
104
abstract
stringlengths
6
4.09k
2501.03830
MeshConv3D: Efficient convolution and pooling operators for triangular 3D meshes
cs.CV cs.GR
Convolutional neural networks (CNNs) have been pivotal in various 2D image analysis tasks, including computer vision, image indexing and retrieval or semantic classification. Extending CNNs to 3D data such as point clouds and 3D meshes raises significant challenges since the very basic convolution and pooling operators need to be completely re-visited and re-defined in an appropriate manner to tackle irregular connectivity issues. In this paper, we introduce MeshConv3D, a 3D mesh-dedicated methodology integrating specialized convolution and face collapse-based pooling operators. MeshConv3D operates directly on meshes of arbitrary topology, without any need of prior re-meshing/conversion techniques. In order to validate our approach, we have considered a semantic classification task. The experimental results obtained on three distinct benchmark datasets show that the proposed approach makes it possible to achieve equivalent or superior classification results, while minimizing the related memory footprint and computational load.
2501.03832
Three-dimensional attention Transformer for state evaluation in real-time strategy games
cs.LG cs.AI
Situation assessment in Real-Time Strategy (RTS) games is crucial for understanding decision-making in complex adversarial environments. However, existing methods remain limited in processing multi-dimensional feature information and temporal dependencies. Here we propose a tri-dimensional Space-Time-Feature Transformer (TSTF Transformer) architecture, which efficiently models battlefield situations through three independent but cascaded modules: spatial attention, temporal attention, and feature attention. On a dataset comprising 3,150 adversarial experiments, the 8-layer TSTF Transformer demonstrates superior performance: achieving 58.7% accuracy in the early game (~4% progress), significantly outperforming the conventional Timesformer's 41.8%; reaching 97.6% accuracy in the mid-game (~40% progress) while maintaining low performance variation (standard deviation 0.114). Meanwhile, this architecture requires fewer parameters (4.75M) compared to the baseline model (5.54M). Our study not only provides new insights into situation assessment in RTS games but also presents an innovative paradigm for Transformer-based multi-dimensional temporal modeling.
2501.03833
Sequence Reconstruction for the Single-Deletion Single-Substitution Channel
cs.IT math.IT
The central problem in sequence reconstruction is to find the minimum number of distinct channel outputs required to uniquely reconstruct the transmitted sequence. According to Levenshtein's work in 2001, this number is determined by the size of the maximum intersection between the error balls of any two distinct input sequences of the channel. In this work, we study the sequence reconstruction problem for single-deletion single-substitution channel, assuming that the transmitted sequence belongs to a $q$-ary code with minimum Hamming distance at least $2$, where $q\geq 2$ is any fixed integer. Specifically, we prove that for any two $q$-ary sequences of length $n$ and with Hamming distance $d\geq 2$, the size of the intersection of their error balls is upper bounded by $2qn-3q-2-\delta_{q,2}$, where $\delta_{i,j}$ is the Kronecker delta. We also prove the tightness of this bound by constructing two sequences the intersection size of whose error balls achieves this bound.
2501.03835
TACLR: A Scalable and Efficient Retrieval-based Method for Industrial Product Attribute Value Identification
cs.CL cs.AI cs.IR
Product Attribute Value Identification (PAVI) involves identifying attribute values from product profiles, a key task for improving product search, recommendations, and business analytics on e-commerce platforms. However, existing PAVI methods face critical challenges, such as inferring implicit values, handling out-of-distribution (OOD) values, and producing normalized outputs. To address these limitations, we introduce Taxonomy-Aware Contrastive Learning Retrieval (TACLR), the first retrieval-based method for PAVI. TACLR formulates PAVI as an information retrieval task by encoding product profiles and candidate values into embeddings and retrieving values based on their similarity to the item embedding. It leverages contrastive training with taxonomy-aware hard negative sampling and employs adaptive inference with dynamic thresholds. TACLR offers three key advantages: (1) it effectively handles implicit and OOD values while producing normalized outputs; (2) it scales to thousands of categories, tens of thousands of attributes, and millions of values; and (3) it supports efficient inference for high-load industrial scenarios. Extensive experiments on proprietary and public datasets validate the effectiveness and efficiency of TACLR. Moreover, it has been successfully deployed in a real-world e-commerce platform, processing millions of product listings daily while supporting dynamic, large-scale attribute taxonomies.
2501.03836
SCC-YOLO: An Improved Object Detector for Assisting in Brain Tumor Diagnosis
eess.IV cs.AI cs.CV
Brain tumors can result in neurological dysfunction, alterations in cognitive and psychological states, increased intracranial pressure, and the occurrence of seizures, thereby presenting a substantial risk to human life and health. The You Only Look Once(YOLO) series models have demonstrated superior accuracy in object detection for medical imaging. In this paper, we develop a novel SCC-YOLO architecture by integrating the SCConv attention mechanism into YOLOv9. The SCConv module reconstructs an efficient convolutional module by reducing spatial and channel redundancy among features, thereby enhancing the learning of image features. We investigate the impact of intergrating different attention mechanisms with the YOLOv9 model on brain tumor image detection using both the Br35H dataset and our self-made dataset(Brain_Tumor_Dataset). Experimental results show that on the Br35H dataset, SCC-YOLO achieved a 0.3% improvement in mAp50 compared to YOLOv9, while on our self-made dataset, SCC-YOLO exhibited a 0.5% improvement over YOLOv9. SCC-YOLO has reached state-of-the-art performance in brain tumor detection. Source code is available at : https://jihulab.com/healthcare-information-studio/SCC-YOLO/-/tree/master
2501.03838
LM-Net: A Light-weight and Multi-scale Network for Medical Image Segmentation
cs.CV
Current medical image segmentation approaches have limitations in deeply exploring multi-scale information and effectively combining local detail textures with global contextual semantic information. This results in over-segmentation, under-segmentation, and blurred segmentation boundaries. To tackle these challenges, we explore multi-scale feature representations from different perspectives, proposing a novel, lightweight, and multi-scale architecture (LM-Net) that integrates advantages of both Convolutional Neural Networks (CNNs) and Vision Transformers (ViTs) to enhance segmentation accuracy. LM-Net employs a lightweight multi-branch module to capture multi-scale features at the same level. Furthermore, we introduce two modules to concurrently capture local detail textures and global semantics with multi-scale features at different levels: the Local Feature Transformer (LFT) and Global Feature Transformer (GFT). The LFT integrates local window self-attention to capture local detail textures, while the GFT leverages global self-attention to capture global contextual semantics. By combining these modules, our model achieves complementarity between local and global representations, alleviating the problem of blurred segmentation boundaries in medical image segmentation. To evaluate the feasibility of LM-Net, extensive experiments have been conducted on three publicly available datasets with different modalities. Our proposed model achieves state-of-the-art results, surpassing previous methods, while only requiring 4.66G FLOPs and 5.4M parameters. These state-of-the-art results on three datasets with different modalities demonstrate the effectiveness and adaptability of our proposed LM-Net for various medical image segmentation tasks.
2501.03839
MedFocusCLIP : Improving few shot classification in medical datasets using pixel wise attention
eess.IV cs.CV
With the popularity of foundational models, parameter efficient fine tuning has become the defacto approach to leverage pretrained models to perform downstream tasks. Taking inspiration from recent advances in large language models, Visual Prompt Tuning, and similar techniques, learn an additional prompt to efficiently finetune a pretrained vision foundational model. However, we observe that such prompting is insufficient for fine-grained visual classification tasks such as medical image classification, where there is large inter-class variance, and small intra-class variance. Hence, in this paper we propose to leverage advanced segmentation capabilities of Segment Anything Model 2 (SAM2) as a visual prompting cue to help visual encoder in the CLIP (Contrastive Language-Image Pretraining) by guiding the attention in CLIP visual encoder to relevant regions in the image. This helps the model to focus on highly discriminative regions, without getting distracted from visually similar background features, an essential requirement in a fewshot, finegrained classification setting. We evaluate our method on diverse medical datasets including X-rays, CT scans, and MRI images, and report an accuracy of (71%, 81%, 86%, 58%) from the proposed approach on (COVID, lung-disease, brain-tumor, breast-cancer) datasets against (66%, 70%, 68%, 29%) from a pretrained CLIP model after fewshot training. The proposed approach also allows to obtain interpretable explanation for the classification performance through the localization obtained using segmentation.
2501.03840
Machine learning applications in archaeological practices: a review
cs.LG
Artificial intelligence and machine learning applications in archaeology have increased significantly in recent years, and these now span all subfields, geographical regions, and time periods. The prevalence and success of these applications have remained largely unexamined, as recent reviews on the use of machine learning in archaeology have only focused only on specific subfields of archaeology. Our review examined an exhaustive corpus of 135 articles published between 1997 and 2022. We observed a significant increase in the number of publications from 2019 onwards. Automatic structure detection and artefact classification were the most represented tasks in the articles reviewed, followed by taphonomy, and archaeological predictive modelling. From the review, clustering and unsupervised methods were underrepresented compared to supervised models. Artificial neural networks and ensemble learning account for two thirds of the total number of models used. However, if machine learning models are gaining in popularity they remain subject to misunderstanding. We observed, in some cases, poorly defined requirements and caveats of the machine learning methods used. Furthermore, the goals and the needs of machine learning applications for archaeological purposes are in some cases unclear or poorly expressed. To address this, we proposed a workflow guide for archaeologists to develop coherent and consistent methodologies adapted to their research questions, project scale and data. As in many other areas, machine learning is rapidly becoming an important tool in archaeological research and practice, useful for the analyses of large and multivariate data, although not without limitations. This review highlights the importance of well-defined and well-reported structured methodologies and collaborative practices to maximise the potential of applications of machine learning methods in archaeology.
2501.03841
OmniManip: Towards General Robotic Manipulation via Object-Centric Interaction Primitives as Spatial Constraints
cs.RO
The development of general robotic systems capable of manipulating in unstructured environments is a significant challenge. While Vision-Language Models(VLM) excel in high-level commonsense reasoning, they lack the fine-grained 3D spatial understanding required for precise manipulation tasks. Fine-tuning VLM on robotic datasets to create Vision-Language-Action Models(VLA) is a potential solution, but it is hindered by high data collection costs and generalization issues. To address these challenges, we propose a novel object-centric representation that bridges the gap between VLM's high-level reasoning and the low-level precision required for manipulation. Our key insight is that an object's canonical space, defined by its functional affordances, provides a structured and semantically meaningful way to describe interaction primitives, such as points and directions. These primitives act as a bridge, translating VLM's commonsense reasoning into actionable 3D spatial constraints. In this context, we introduce a dual closed-loop, open-vocabulary robotic manipulation system: one loop for high-level planning through primitive resampling, interaction rendering and VLM checking, and another for low-level execution via 6D pose tracking. This design ensures robust, real-time control without requiring VLM fine-tuning. Extensive experiments demonstrate strong zero-shot generalization across diverse robotic manipulation tasks, highlighting the potential of this approach for automating large-scale simulation data generation.
2501.03843
BERTopic for Topic Modeling of Hindi Short Texts: A Comparative Study
cs.IR cs.CL cs.LG
As short text data in native languages like Hindi increasingly appear in modern media, robust methods for topic modeling on such data have gained importance. This study investigates the performance of BERTopic in modeling Hindi short texts, an area that has been under-explored in existing research. Using contextual embeddings, BERTopic can capture semantic relationships in data, making it potentially more effective than traditional models, especially for short and diverse texts. We evaluate BERTopic using 6 different document embedding models and compare its performance against 8 established topic modeling techniques, such as Latent Dirichlet Allocation (LDA), Non-negative Matrix Factorization (NMF), Latent Semantic Indexing (LSI), Additive Regularization of Topic Models (ARTM), Probabilistic Latent Semantic Analysis (PLSA), Embedded Topic Model (ETM), Combined Topic Model (CTM), and Top2Vec. The models are assessed using coherence scores across a range of topic counts. Our results reveal that BERTopic consistently outperforms other models in capturing coherent topics from short Hindi texts.
2501.03847
Diffusion as Shader: 3D-aware Video Diffusion for Versatile Video Generation Control
cs.CV cs.AI cs.GR
Diffusion models have demonstrated impressive performance in generating high-quality videos from text prompts or images. However, precise control over the video generation process, such as camera manipulation or content editing, remains a significant challenge. Existing methods for controlled video generation are typically limited to a single control type, lacking the flexibility to handle diverse control demands. In this paper, we introduce Diffusion as Shader (DaS), a novel approach that supports multiple video control tasks within a unified architecture. Our key insight is that achieving versatile video control necessitates leveraging 3D control signals, as videos are fundamentally 2D renderings of dynamic 3D content. Unlike prior methods limited to 2D control signals, DaS leverages 3D tracking videos as control inputs, making the video diffusion process inherently 3D-aware. This innovation allows DaS to achieve a wide range of video controls by simply manipulating the 3D tracking videos. A further advantage of using 3D tracking videos is their ability to effectively link frames, significantly enhancing the temporal consistency of the generated videos. With just 3 days of fine-tuning on 8 H800 GPUs using less than 10k videos, DaS demonstrates strong control capabilities across diverse tasks, including mesh-to-video generation, camera control, motion transfer, and object manipulation.
2501.03848
Semise: Semi-supervised learning for severity representation in medical image
eess.IV cs.CV
This paper introduces SEMISE, a novel method for representation learning in medical imaging that combines self-supervised and supervised learning. By leveraging both labeled and augmented data, SEMISE addresses the challenge of data scarcity and enhances the encoder's ability to extract meaningful features. This integrated approach leads to more informative representations, improving performance on downstream tasks. As result, our approach achieved a 12% improvement in classification and a 3% improvement in segmentation, outperforming existing methods. These results demonstrate the potential of SIMESE to advance medical image analysis and offer more accurate solutions for healthcare applications, particularly in contexts where labeled data is limited.
2501.03850
Partitioning Strategies for Parallel Computation of Flexible Skylines
cs.DB
While classical skyline queries identify interesting data within large datasets, flexible skylines introduce preferences through constraints on attribute weights, and further reduce the data returned. However, computing these queries can be time-consuming for large datasets. We propose and implement a parallel computation scheme consisting of a parallel phase followed by a sequential phase, and apply it to flexible skylines. We assess the additional effect of an initial filtering phase to reduce dataset size before parallel processing, and the elimination of the sequential part (the most time-consuming) altogether. All our experiments are executed in the PySpark framework for a number of different datasets of varying sizes and dimensions.
2501.03853
Leveraging time and parameters for nonlinear model reduction methods
math.NA cs.LG cs.NA
In this paper, we consider model order reduction (MOR) methods for problems with slowly decaying Kolmogorov $n$-widths as, e.g., certain wave-like or transport-dominated problems. To overcome this Kolmogorov barrier within MOR, nonlinear projections are used, which are often realized numerically using autoencoders. These autoencoders generally consist of a nonlinear encoder and a nonlinear decoder and involve costly training of the hyperparameters to obtain a good approximation quality of the reduced system. To facilitate the training process, we show that extending the to-be-reduced system and its corresponding training data makes it possible to replace the nonlinear encoder with a linear encoder without sacrificing accuracy, thus roughly halving the number of hyperparameters to be trained.
2501.03854
Comparison of Integration Methods for Cut Elements
cs.CE
Using an interface inserted in a background mesh is an alternative way of constructing a complex geometrical shape with a relative low meshing efforts. However, this process may require special treatment of elements cut by the interface. Our study focuses on comparing the integration of cut elements defined by implicit and parametric curves. We investigate the efficiency and robustness of open-source tools such as Algoim [5](a library for quadrature on implicitly defined geometries) and Ginkgo [2](a library for isogeometric analysis on Boolean operations with a parametric description) with numerical examples computing the area defined by the interface and benchmarks for 2D elasticity problem using the open-source code GeoPDEs [7]. It is concluded that none of the two interface descriptions is preferable with respect to the quality of the integration. Thus, the choice of the interface type depends only on the studied problem and the available curve description, but not on the numerical aspects of the integration.
2501.03855
BabyLMs for isiXhosa: Data-Efficient Language Modelling in a Low-Resource Context
cs.CL
The BabyLM challenge called on participants to develop sample-efficient language models. Submissions were pretrained on a fixed English corpus, limited to the amount of words children are exposed to in development (<100m). The challenge produced new architectures for data-efficient language modelling, which outperformed models trained on trillions of words. This is promising for low-resource languages, where available corpora are limited to much less than 100m words. In this paper, we explore the potential of BabyLMs for low-resource languages, using the isiXhosa language as a case study. We pretrain two BabyLM architectures, ELC-BERT and MLSM, on an isiXhosa corpus. They outperform a vanilla pretrained model on POS tagging and NER, achieving notable gains (+3.2 F1) for the latter. In some instances, the BabyLMs even outperform XLM-R. Our findings show that data-efficient models are viable for low-resource languages, but highlight the continued importance, and lack of, high-quality pretraining data. Finally, we visually analyse how BabyLM architectures encode isiXhosa.
2501.03857
Progressive Document-level Text Simplification via Large Language Models
cs.CL
Research on text simplification has primarily focused on lexical and sentence-level changes. Long document-level simplification (DS) is still relatively unexplored. Large Language Models (LLMs), like ChatGPT, have excelled in many natural language processing tasks. However, their performance on DS tasks is unsatisfactory, as they often treat DS as merely document summarization. For the DS task, the generated long sequences not only must maintain consistency with the original document throughout, but complete moderate simplification operations encompassing discourses, sentences, and word-level simplifications. Human editors employ a hierarchical complexity simplification strategy to simplify documents. This study delves into simulating this strategy through the utilization of a multi-stage collaboration using LLMs. We propose a progressive simplification method (ProgDS) by hierarchically decomposing the task, including the discourse-level, topic-level, and lexical-level simplification. Experimental results demonstrate that ProgDS significantly outperforms existing smaller models or direct prompting with LLMs, advancing the state-of-the-art in the document simplification task.
2501.03858
Symmetry and Generalisation in Machine Learning
cs.LG stat.ML
This work is about understanding the impact of invariance and equivariance on generalisation in supervised learning. We use the perspective afforded by an averaging operator to show that for any predictor that is not equivariant, there is an equivariant predictor with strictly lower test risk on all regression problems where the equivariance is correctly specified. This constitutes a rigorous proof that symmetry, in the form of invariance or equivariance, is a useful inductive bias. We apply these ideas to equivariance and invariance in random design least squares and kernel ridge regression respectively. This allows us to specify the reduction in expected test risk in more concrete settings and express it in terms of properties of the group, the model and the data. Along the way, we give examples and additional results to demonstrate the utility of the averaging operator approach in analysing equivariant predictors. In addition, we adopt an alternative perspective and formalise the common intuition that learning with invariant models reduces to a problem in terms of orbit representatives. The formalism extends naturally to a similar intuition for equivariant models. We conclude by connecting the two perspectives and giving some ideas for future work.
2501.03859
A Synergistic Framework for Learning Shape Estimation and Shape-Aware Whole-Body Control Policy for Continuum Robots
cs.RO
In this paper, we present a novel synergistic framework for learning shape estimation and a shape-aware whole-body control policy for tendon-driven continuum robots. Our approach leverages the interaction between two Augmented Neural Ordinary Differential Equations (ANODEs) -- the Shape-NODE and Control-NODE -- to achieve continuous shape estimation and shape-aware control. The Shape-NODE integrates prior knowledge from Cosserat rod theory, allowing it to adapt and account for model mismatches, while the Control-NODE uses this shape information to optimize a whole-body control policy, trained in a Model Predictive Control (MPC) fashion. This unified framework effectively overcomes limitations of existing data-driven methods, such as poor shape awareness and challenges in capturing complex nonlinear dynamics. Extensive evaluations in both simulation and real-world environments demonstrate the framework's robust performance in shape estimation, trajectory tracking, and obstacle avoidance. The proposed method consistently outperforms state-of-the-art end-to-end, Neural-ODE, and Recurrent Neural Network (RNN) models, particularly in terms of tracking accuracy and generalization capabilities.
2501.03863
Improving Dialectal Slot and Intent Detection with Auxiliary Tasks: A Multi-Dialectal Bavarian Case Study
cs.CL
Reliable slot and intent detection (SID) is crucial in natural language understanding for applications like digital assistants. Encoder-only transformer models fine-tuned on high-resource languages generally perform well on SID. However, they struggle with dialectal data, where no standardized form exists and training data is scarce and costly to produce. We explore zero-shot transfer learning for SID, focusing on multiple Bavarian dialects, for which we release a new dataset for the Munich dialect. We evaluate models trained on auxiliary tasks in Bavarian, and compare joint multi-task learning with intermediate-task training. We also compare three types of auxiliary tasks: token-level syntactic tasks, named entity recognition (NER), and language modelling. We find that the included auxiliary tasks have a more positive effect on slot filling than intent classification (with NER having the most positive effect), and that intermediate-task training yields more consistent performance gains. Our best-performing approach improves intent classification performance on Bavarian dialects by 5.1 and slot filling F1 by 8.4 percentage points.
2501.03865
Truthful mechanisms for linear bandit games with private contexts
cs.LG cs.GT
The contextual bandit problem, where agents arrive sequentially with personal contexts and the system adapts its arm allocation decisions accordingly, has recently garnered increasing attention for enabling more personalized outcomes. However, in many healthcare and recommendation applications, agents have private profiles and may misreport their contexts to gain from the system. For example, in adaptive clinical trials, where hospitals sequentially recruit volunteers to test multiple new treatments and adjust plans based on volunteers' reported profiles such as symptoms and interim data, participants may misreport severe side effects like allergy and nausea to avoid perceived suboptimal treatments. We are the first to study this issue of private context misreporting in a stochastic contextual bandit game between the system and non-repeated agents. We show that traditional low-regret algorithms, such as UCB family algorithms and Thompson sampling, fail to ensure truthful reporting and can result in linear regret in the worst case, while traditional truthful algorithms like explore-then-commit (ETC) and $\epsilon$-greedy algorithm incur sublinear but high regret. We propose a mechanism that uses a linear program to ensure truthfulness while minimizing deviation from Thompson sampling, yielding an $O(\ln T)$ frequentist regret. Our numerical experiments further demonstrate strong performance in multiple contexts and across other distribution families.
2501.03870
Add Noise, Tasks, or Layers? MaiNLP at the VarDial 2025 Shared Task on Norwegian Dialectal Slot and Intent Detection
cs.CL
Slot and intent detection (SID) is a classic natural language understanding task. Despite this, research has only more recently begun focusing on SID for dialectal and colloquial varieties. Many approaches for low-resource scenarios have not yet been applied to dialectal SID data, or compared to each other on the same datasets. We participate in the VarDial 2025 shared task on slot and intent detection in Norwegian varieties, and compare multiple set-ups: varying the training data (English, Norwegian, or dialectal Norwegian), injecting character-level noise, training on auxiliary tasks, and applying Layer Swapping, a technique in which layers of models fine-tuned on different datasets are assembled into a model. We find noise injection to be beneficial while the effects of auxiliary tasks are mixed. Though some experimentation was required to successfully assemble a model from layers, it worked surprisingly well; a combination of models trained on English and small amounts of dialectal data produced the most robust slot predictions. Our best models achieve 97.6% intent accuracy and 85.6% slot F1 in the shared task.
2501.03874
Neuromorphic Optical Tracking and Imaging of Randomly Moving Targets through Strongly Scattering Media
cs.NE cs.CV cs.LG eess.IV
Tracking and acquiring simultaneous optical images of randomly moving targets obscured by scattering media remains a challenging problem of importance to many applications that require precise object localization and identification. In this work we develop an end-to-end neuromorphic optical engineering and computational approach to demonstrate how to track and image normally invisible objects by combining an event detecting camera with a multistage neuromorphic deep learning strategy. Photons emerging from dense scattering media are detected by the event camera and converted to pixel-wise asynchronized spike trains - a first step in isolating object-specific information from the dominant uninformative background. Spiking data is fed into a deep spiking neural network (SNN) engine where object tracking and image reconstruction are performed by two separate yet interconnected modules running in parallel in discrete time steps over the event duration. Through benchtop experiments we demonstrate tracking and imaging randomly moving objects in dense turbid media as well as image reconstruction of spatially stationary but optically dynamic objects. Standardized character sets serve as representative proxies for geometrically complex objects, underscoring the method's generality. The results highlight the advantages of a fully neuromorphic approach in meeting a major imaging technology with high computational efficiency and low power consumption.
2501.03875
ZDySS -- Zero-Shot Dynamic Scene Stylization using Gaussian Splatting
cs.CV
Stylizing a dynamic scene based on an exemplar image is critical for various real-world applications, including gaming, filmmaking, and augmented and virtual reality. However, achieving consistent stylization across both spatial and temporal dimensions remains a significant challenge. Most existing methods are designed for static scenes and often require an optimization process for each style image, limiting their adaptability. We introduce ZDySS, a zero-shot stylization framework for dynamic scenes, allowing our model to generalize to previously unseen style images at inference. Our approach employs Gaussian splatting for scene representation, linking each Gaussian to a learned feature vector that renders a feature map for any given view and timestamp. By applying style transfer on the learned feature vectors instead of the rendered feature map, we enhance spatio-temporal consistency across frames. Our method demonstrates superior performance and coherence over state-of-the-art baselines in tests on real-world dynamic scenes, making it a robust solution for practical applications.
2501.03877
Stochastically Constrained Best Arm Identification with Thompson Sampling
cs.LG
We consider the problem of the best arm identification in the presence of stochastic constraints, where there is a finite number of arms associated with multiple performance measures. The goal is to identify the arm that optimizes the objective measure subject to constraints on the remaining measures. We will explore the popular idea of Thompson sampling (TS) as a means to solve it. To the best of our knowledge, it is the first attempt to extend TS to this problem. We will design a TS-based sampling algorithm, establish its asymptotic optimality in the rate of posterior convergence, and demonstrate its superior performance using numerical examples.
2501.03879
CL3DOR: Contrastive Learning for 3D Large Multimodal Models via Odds Ratio on High-Resolution Point Clouds
cs.CV cs.AI
Recent research has demonstrated that Large Language Models (LLMs) are not limited to text-only tasks but can also function as multimodal models across various modalities, including audio, images, and videos. In particular, research on 3D Large Multimodal Models (3D LMMs) is making notable strides, driven by the potential of processing higher-dimensional data like point clouds. However, upon closer examination, we find that the visual and textual content within each sample of existing training datasets lacks both high informational granularity and clarity, which serve as a bottleneck for precise cross-modal understanding. To address these issues, we propose CL3DOR, Contrastive Learning for 3D large multimodal models via Odds ratio on high-Resolution point clouds, designed to ensure greater specificity and clarity in both visual and textual content. Specifically, we increase the density of point clouds per object and construct informative hard negative responses in the training dataset to penalize unwanted responses. To leverage hard negative responses, we incorporate the odds ratio as an auxiliary term for contrastive learning into the conventional language modeling loss. CL3DOR achieves state-of-the-art performance in 3D scene understanding and reasoning benchmarks. Additionally, we demonstrate the effectiveness of CL3DOR's key components through extensive experiments.
2501.03880
SELMA3D challenge: Self-supervised learning for 3D light-sheet microscopy image segmentation
eess.IV cs.CV cs.LG
Recent innovations in light sheet microscopy, paired with developments in tissue clearing techniques, enable the 3D imaging of large mammalian tissues with cellular resolution. Combined with the progress in large-scale data analysis, driven by deep learning, these innovations empower researchers to rapidly investigate the morphological and functional properties of diverse biological samples. Segmentation, a crucial preliminary step in the analysis process, can be automated using domain-specific deep learning models with expert-level performance. However, these models exhibit high sensitivity to domain shifts, leading to a significant drop in accuracy when applied to data outside their training distribution. To address this limitation, and inspired by the recent success of self-supervised learning in training generalizable models, we organized the SELMA3D Challenge during the MICCAI 2024 conference. SELMA3D provides a vast collection of light-sheet images from cleared mice and human brains, comprising 35 large 3D images-each with over 1000^3 voxels-and 315 annotated small patches for finetuning, preliminary testing and final testing. The dataset encompasses diverse biological structures, including vessel-like and spot-like structures. Five teams participated in all phases of the challenge, and their proposed methods are reviewed in this paper. Quantitative and qualitative results from most participating teams demonstrate that self-supervised learning on large datasets improves segmentation model performance and generalization. We will continue to support and extend SELMA3D as an inaugural MICCAI challenge focused on self-supervised learning for 3D microscopy image segmentation.
2501.03881
An LSTM-based Test Selection Method for Self-Driving Cars
cs.RO cs.SE
Self-driving cars require extensive testing, which can be costly in terms of time. To optimize this process, simple and straightforward tests should be excluded, focusing on challenging tests instead. This study addresses the test selection problem for lane-keeping systems for self-driving cars. Road segment features, such as angles and lengths, were extracted and treated as sequences, enabling classification of the test cases as "safe" or "unsafe" using a long short-term memory (LSTM) model. The proposed model is compared against machine learning-based test selectors. Results demonstrated that the LSTM-based method outperformed machine learning-based methods in accuracy and precision metrics while exhibiting comparable performance in recall and F1 scores. This work introduces a novel deep learning-based approach to the road classification problem, providing an effective solution for self-driving car test selection using a simulation environment.
2501.03884
AlphaPO - Reward shape matters for LLM alignment
cs.CL
Reinforcement Learning with Human Feedback (RLHF) and its variants have made huge strides toward the effective alignment of large language models (LLMs) to follow instructions and reflect human values. More recently, Direct Alignment Algorithms (DAAs) have emerged in which the reward modeling stage of RLHF is skipped by characterizing the reward directly as a function of the policy being learned. Some popular examples of DAAs include Direct Preference Optimization (DPO) and Simple Preference Optimization (SimPO). These methods often suffer from likelihood displacement, a phenomenon by which the probabilities of preferred responses are often reduced undesirably. In this paper, we argue that, for DAAs the reward (function) shape matters. We introduce \textbf{AlphaPO}, a new DAA method that leverages an $\alpha$-parameter to help change the shape of the reward function beyond the standard log reward. AlphaPO helps maintain fine-grained control over likelihood displacement and over-optimization. Compared to SimPO, one of the best performing DAAs, AlphaPO leads to about 7\% to 10\% relative improvement in alignment performance for the instruct versions of Mistral-7B and Llama3-8B while achieving 15\% to 50\% relative improvement over DPO on the same models. The analysis and results presented highlight the importance of the reward shape, and how one can systematically change it to affect training dynamics, as well as improve alignment performance.
2501.03888
Neural DNF-MT: A Neuro-symbolic Approach for Learning Interpretable and Editable Policies
cs.AI cs.LG cs.LO
Although deep reinforcement learning has been shown to be effective, the model's black-box nature presents barriers to direct policy interpretation. To address this problem, we propose a neuro-symbolic approach called neural DNF-MT for end-to-end policy learning. The differentiable nature of the neural DNF-MT model enables the use of deep actor-critic algorithms for training. At the same time, its architecture is designed so that trained models can be directly translated into interpretable policies expressed as standard (bivalent or probabilistic) logic programs. Moreover, additional layers can be included to extract abstract features from complex observations, acting as a form of predicate invention. The logic representations are highly interpretable, and we show how the bivalent representations of deterministic policies can be edited and incorporated back into a neural model, facilitating manual intervention and adaptation of learned policies. We evaluate our approach on a range of tasks requiring learning deterministic or stochastic behaviours from various forms of observations. Our empirical results show that our neural DNF-MT model performs at the level of competing black-box methods whilst providing interpretable policies.
2501.03891
Superpixel Boundary Correction for Weakly-Supervised Semantic Segmentation on Histopathology Images
cs.CV
With the rapid advancement of deep learning, computational pathology has made significant progress in cancer diagnosis and subtyping. Tissue segmentation is a core challenge, essential for prognosis and treatment decisions. Weakly supervised semantic segmentation (WSSS) reduces the annotation requirement by using image-level labels instead of pixel-level ones. However, Class Activation Map (CAM)-based methods still suffer from low spatial resolution and unclear boundaries. To address these issues, we propose a multi-level superpixel correction algorithm that refines CAM boundaries using superpixel clustering and floodfill. Experimental results show that our method achieves great performance on breast cancer segmentation dataset with mIoU of 71.08%, significantly improving tumor microenvironment boundary delineation.
2501.03892
LEAP: LLM-powered End-to-end Automatic Library for Processing Social Science Queries on Unstructured Data
cs.DB
Social scientists are increasingly interested in analyzing the semantic information (e.g., emotion) of unstructured data (e.g., Tweets), where the semantic information is not natively present. Performing this analysis in a cost-efficient manner requires using machine learning (ML) models to extract the semantic information and subsequently analyze the now structured data. However, this process remains challenging for domain experts. To demonstrate the challenges in social science analytics, we collect a dataset, QUIET-ML, of 120 real-world social science queries in natural language and their ground truth answers. Existing systems struggle with these queries since (1) they require selecting and applying ML models, and (2) more than a quarter of these queries are vague, making standard tools like natural language to SQL systems unsuited. To address these issues, we develop LEAP, an end-to-end library that answers social science queries in natural language with ML. LEAP filters vague queries to ensure that the answers are deterministic and selects from internally supported and user-defined ML functions to extend the unstructured data to structured tables with necessary annotations. LEAP further generates and executes code to respond to these natural language queries. LEAP achieves a 100% pass @ 3 and 92% pass @ 1 on QUIET-ML, with a \$1.06 average end-to-end cost, of which code generation costs \$0.02.
2501.03894
Robust Moving-horizon Estimation for Nonlinear Systems: From Perfect to Imperfect Optimization
eess.SY cs.SY
Robust stability of moving-horizon estimators is investigated for nonlinear discrete-time systems that are detectable in the sense of incremental input/output-to-state stability and are affected by disturbances. The estimate of a moving-horizon estimator stems from the on-line solution of a least-squares minimization problem at each time instant. The resulting stability guarantees depend on the optimization tolerance in solving such minimization problems. Specifically, two main contributions are established: (i) the robust stability of the estimation error, while supposing to solve exactly the on-line minimization problem; (ii) the practical robust stability of the estimation error with state estimates obtained by an imperfect minimization. Finally, the construction of such robust moving-horizon estimators and the performances resulting from the design based on the theoretical findings are showcased with two numerical examples.
2501.03895
LLaVA-Mini: Efficient Image and Video Large Multimodal Models with One Vision Token
cs.CV cs.AI cs.CL
The advent of real-time large multimodal models (LMMs) like GPT-4o has sparked considerable interest in efficient LMMs. LMM frameworks typically encode visual inputs into vision tokens (continuous representations) and integrate them and textual instructions into the context of large language models (LLMs), where large-scale parameters and numerous context tokens (predominantly vision tokens) result in substantial computational overhead. Previous efforts towards efficient LMMs always focus on replacing the LLM backbone with smaller models, while neglecting the crucial issue of token quantity. In this paper, we introduce LLaVA-Mini, an efficient LMM with minimal vision tokens. To achieve a high compression ratio of vision tokens while preserving visual information, we first analyze how LMMs understand vision tokens and find that most vision tokens only play a crucial role in the early layers of LLM backbone, where they mainly fuse visual information into text tokens. Building on this finding, LLaVA-Mini introduces modality pre-fusion to fuse visual information into text tokens in advance, thereby facilitating the extreme compression of vision tokens fed to LLM backbone into one token. LLaVA-Mini is a unified large multimodal model that can support the understanding of images, high-resolution images, and videos in an efficient manner. Experiments across 11 image-based and 7 video-based benchmarks demonstrate that LLaVA-Mini outperforms LLaVA-v1.5 with just 1 vision token instead of 576. Efficiency analyses reveal that LLaVA-Mini can reduce FLOPs by 77%, deliver low-latency responses within 40 milliseconds, and process over 10,000 frames of video on the GPU hardware with 24GB of memory.
2501.03902
Explainable Reinforcement Learning via Temporal Policy Decomposition
cs.LG cs.AI
We investigate the explainability of Reinforcement Learning (RL) policies from a temporal perspective, focusing on the sequence of future outcomes associated with individual actions. In RL, value functions compress information about rewards collected across multiple trajectories and over an infinite horizon, allowing a compact form of knowledge representation. However, this compression obscures the temporal details inherent in sequential decision-making, presenting a key challenge for interpretability. We present Temporal Policy Decomposition (TPD), a novel explainability approach that explains individual RL actions in terms of their Expected Future Outcome (EFO). These explanations decompose generalized value functions into a sequence of EFOs, one for each time step up to a prediction horizon of interest, revealing insights into when specific outcomes are expected to occur. We leverage fixed-horizon temporal difference learning to devise an off-policy method for learning EFOs for both optimal and suboptimal actions, enabling contrastive explanations consisting of EFOs for different state-action pairs. Our experiments demonstrate that TPD generates accurate explanations that (i) clarify the policy's future strategy and anticipated trajectory for a given action and (ii) improve understanding of the reward composition, facilitating fine-tuning of the reward function to align with human expectations.
2501.03904
Exploring the Potential of Large Language Models in Public Transportation: San Antonio Case Study
cs.LG cs.AI cs.IR
The integration of large language models (LLMs) into public transit systems presents a transformative opportunity to enhance urban mobility. This study explores the potential of LLMs to revolutionize public transportation management within the context of San Antonio's transit system. Leveraging the capabilities of LLMs in natural language processing and data analysis, we investigate their capabilities to optimize route planning, reduce wait times, and provide personalized travel assistance. By utilizing the General Transit Feed Specification (GTFS) and other relevant data, this research aims to demonstrate how LLMs can potentially improve resource allocation, elevate passenger satisfaction, and inform data-driven decision-making in transit operations. A comparative analysis of different ChatGPT models was conducted to assess their ability to understand transportation information, retrieve relevant data, and provide comprehensive responses. Findings from this study suggest that while LLMs hold immense promise for public transit, careful engineering and fine-tuning are essential to realizing their full potential. San Antonio serves as a case study to inform the development of LLM-powered transit systems in other urban environments.
2501.03905
mFabric: An Efficient and Scalable Fabric for Mixture-of-Experts Training
cs.NI cs.LG
Mixture-of-Expert (MoE) models outperform conventional models by selectively activating different subnets, named \emph{experts}, on a per-token basis. This gated computation generates dynamic communications that cannot be determined beforehand, challenging the existing GPU interconnects that remain \emph{static} during the distributed training process. In this paper, we advocate for a first-of-its-kind system, called mFabric, that unlocks topology reconfiguration \emph{during} distributed MoE training. Towards this vision, we first perform a production measurement study and show that the MoE dynamic communication pattern has \emph{strong locality}, alleviating the requirement of global reconfiguration. Based on this, we design and implement a \emph{regionally reconfigurable high-bandwidth domain} on top of existing electrical interconnects using optical circuit switching (OCS), achieving scalability while maintaining rapid adaptability. We have built a fully functional mFabric prototype with commodity hardware and a customized collective communication runtime that trains state-of-the-art MoE models with \emph{in-training} topology reconfiguration across 32 A100 GPUs. Large-scale packet-level simulations show that mFabric delivers comparable performance as the non-blocking fat-tree fabric while boosting the training cost efficiency (e.g., performance per dollar) of four representative MoE models by 1.2$\times$--1.5$\times$ and 1.9$\times$--2.3$\times$ at 100 Gbps and 400 Gbps link bandwidths, respectively.
2501.03907
Implicit Coordination using Active Epistemic Inference for Multi-Robot Systems
cs.RO
A Multi-robot system (MRS) provides significant advantages for intricate tasks such as environmental monitoring, underwater inspections, and space missions. However, addressing potential communication failures or the lack of communication infrastructure in these fields remains a challenge. A significant portion of MRS research presumes that the system can maintain communication with proximity constraints, but this approach does not solve situations where communication is either non-existent, unreliable, or poses a security risk. Some approaches tackle this issue using predictions about other robots while not communicating, but these methods generally only permit agents to utilize first-order reasoning, which involves reasoning based purely on their own observations. In contrast, to deal with this problem, our proposed framework utilizes Theory of Mind (ToM), employing higher-order reasoning by shifting a robot's perspective to reason about a belief of others observations. Our approach has two main phases: i) an efficient runtime plan adaptation using active inference to signal intentions and reason about a robot's own belief and the beliefs of others in the system, and ii) a hierarchical epistemic planning framework to iteratively reason about the current MRS mission state. The proposed framework outperforms greedy and first-order reasoning approaches and is validated using simulations and experiments with heterogeneous robotic systems.
2501.03910
HYB-VITON: A Hybrid Approach to Virtual Try-On Combining Explicit and Implicit Warping
cs.CV
Virtual try-on systems have significant potential in e-commerce, allowing customers to visualize garments on themselves. Existing image-based methods fall into two categories: those that directly warp garment-images onto person-images (explicit warping), and those using cross-attention to reconstruct given garments (implicit warping). Explicit warping preserves garment details but often produces unrealistic output, while implicit warping achieves natural reconstruction but struggles with fine details. We propose HYB-VITON, a novel approach that combines the advantages of each method and includes both a preprocessing pipeline for warped garments and a novel training option. These components allow us to utilize beneficial regions of explicitly warped garments while leveraging the natural reconstruction of implicit warping. A series of experiments demonstrates that HYB-VITON preserves garment details more faithfully than recent diffusion-based methods, while producing more realistic results than a state-of-the-art explicit warping method.
2501.03916
Dolphin: Closed-loop Open-ended Auto-research through Thinking, Practice, and Feedback
cs.AI cs.CL cs.CV
The scientific research paradigm is undergoing a profound transformation owing to the development of Artificial Intelligence (AI). Recent works demonstrate that various AI-assisted research methods can largely improve research efficiency by improving data analysis, accelerating computation, and fostering novel idea generation. To further move towards the ultimate goal (i.e., automatic scientific research), in this paper, we propose Dolphin, the first closed-loop open-ended auto-research framework to further build the entire process of human scientific research. Dolphin can generate research ideas, perform experiments, and get feedback from experimental results to generate higher-quality ideas. More specifically, Dolphin first generates novel ideas based on relevant papers which are ranked by the topic and task attributes. Then, the codes are automatically generated and debugged with the exception-traceback-guided local code structure. Finally, Dolphin automatically analyzes the results of each idea and feeds the results back to the next round of idea generation. Experiments are conducted on the benchmark datasets of different topics and results show that Dolphin can generate novel ideas continuously and complete the experiment in a loop. We highlight that Dolphin can automatically propose methods that are comparable to the state-of-the-art in some tasks such as 2D image classification and 3D point classification.
2501.03922
Changing almost perfect nonlinear functions on affine subspaces of small codimensions
math.CO cs.IT math.CA math.IT
In this article, we study algebraic decompositions and secondary constructions of almost perfect nonlinear (APN) functions. In many cases, we establish precise criteria which characterize when certain modifications of a given APN function yield new ones. Furthermore, we show that some of the newly constructed functions are extended-affine inequivalent to the original ones.
2501.03923
Explainable AI model reveals disease-related mechanisms in single-cell RNA-seq data
q-bio.GN cs.CV cs.LG
Neurodegenerative diseases (NDDs) are complex and lack effective treatment due to their poorly understood mechanism. The increasingly used data analysis from Single nucleus RNA Sequencing (snRNA-seq) allows to explore transcriptomic events at a single cell level, yet face challenges in interpreting the mechanisms underlying a disease. On the other hand, Neural Network (NN) models can handle complex data to offer insights but can be seen as black boxes with poor interpretability. In this context, explainable AI (XAI) emerges as a solution that could help to understand disease-associated mechanisms when combined with efficient NN models. However, limited research explores XAI in single-cell data. In this work, we implement a method for identifying disease-related genes and the mechanistic explanation of disease progression based on NN model combined with SHAP. We analyze available Huntington's disease (HD) data to identify both HD-altered genes and mechanisms by adding Gene Set Enrichment Analysis (GSEA) comparing two methods, differential gene expression analysis (DGE) and NN combined with SHAP approach. Our results show that DGE and SHAP approaches offer both common and differential sets of altered genes and pathways, reinforcing the usefulness of XAI methods for a broader perspective of disease.
2501.03928
From Newswire to Nexus: Using text-based actor embeddings and transformer networks to forecast conflict dynamics
cs.CY cs.CL cs.LG
This study advances the field of conflict forecasting by using text-based actor embeddings with transformer models to predict dynamic changes in violent conflict patterns at the actor level. More specifically, we combine newswire texts with structured conflict event data and leverage recent advances in Natural Language Processing (NLP) techniques to forecast escalations and de-escalations among conflicting actors, such as governments, militias, separatist movements, and terrorists. This new approach accurately and promptly captures the inherently volatile patterns of violent conflicts, which existing methods have not been able to achieve. To create this framework, we began by curating and annotating a vast international newswire corpus, leveraging hand-labeled event data from the Uppsala Conflict Data Program. By using this hybrid dataset, our models can incorporate the textual context of news sources along with the precision and detail of structured event data. This combination enables us to make both dynamic and granular predictions about conflict developments. We validate our approach through rigorous back-testing against historical events, demonstrating superior out-of-sample predictive power. We find that our approach is quite effective in identifying and predicting phases of conflict escalation and de-escalation, surpassing the capabilities of traditional models. By focusing on actor interactions, our explicit goal is to provide actionable insights to policymakers, humanitarian organizations, and peacekeeping operations in order to enable targeted and effective intervention strategies.
2501.03930
Towards Reliable Testing for Multiple Information Retrieval System Comparisons
cs.IR
Null Hypothesis Significance Testing is the \textit{de facto} tool for assessing effectiveness differences between Information Retrieval systems. Researchers use statistical tests to check whether those differences will generalise to online settings or are just due to the samples observed in the laboratory. Much work has been devoted to studying which test is the most reliable when comparing a pair of systems, but most of the IR real-world experiments involve more than two. In the multiple comparisons scenario, testing several systems simultaneously may inflate the errors committed by the tests. In this paper, we use a new approach to assess the reliability of multiple comparison procedures using simulated and real TREC data. Experiments show that Wilcoxon plus the Benjamini-Hochberg correction yields Type I error rates according to the significance level for typical sample sizes while being the best test in terms of statistical power.
2501.03931
Magic Mirror: ID-Preserved Video Generation in Video Diffusion Transformers
cs.CV
We present Magic Mirror, a framework for generating identity-preserved videos with cinematic-level quality and dynamic motion. While recent advances in video diffusion models have shown impressive capabilities in text-to-video generation, maintaining consistent identity while producing natural motion remains challenging. Previous methods either require person-specific fine-tuning or struggle to balance identity preservation with motion diversity. Built upon Video Diffusion Transformers, our method introduces three key components: (1) a dual-branch facial feature extractor that captures both identity and structural features, (2) a lightweight cross-modal adapter with Conditioned Adaptive Normalization for efficient identity integration, and (3) a two-stage training strategy combining synthetic identity pairs with video data. Extensive experiments demonstrate that Magic Mirror effectively balances identity consistency with natural motion, outperforming existing methods across multiple metrics while requiring minimal parameters added. The code and model will be made publicly available at: https://github.com/dvlab-research/MagicMirror/
2501.03932
CoStruction: Conjoint radiance field optimization for urban scene reconStruction with limited image overlap
cs.CV
Reconstructing the surrounding surface geometry from recorded driving sequences poses a significant challenge due to the limited image overlap and complex topology of urban environments. SoTA neural implicit surface reconstruction methods often struggle in such setting, either failing due to small vision overlap or exhibiting suboptimal performance in accurately reconstructing both the surface and fine structures. To address these limitations, we introduce CoStruction, a novel hybrid implicit surface reconstruction method tailored for large driving sequences with limited camera overlap. CoStruction leverages cross-representation uncertainty estimation to filter out ambiguous geometry caused by limited observations. Our method performs joint optimization of both radiance fields in addition to guided sampling achieving accurate reconstruction of large areas along with fine structures in complex urban scenarios. Extensive evaluation on major driving datasets demonstrates the superiority of our approach in reconstructing large driving sequences with limited image overlap, outperforming concurrent SoTA methods.
2501.03936
PPTAgent: Generating and Evaluating Presentations Beyond Text-to-Slides
cs.AI cs.CL
Automatically generating presentations from documents is a challenging task that requires accommodating content quality, visual appeal, and structural coherence. Existing methods primarily focus on improving and evaluating the content quality in isolation, overlooking visual appeal and structural coherence, which limits their practical applicability. To address these limitations, we propose PPTAgent, which comprehensively improves presentation generation through a two-stage, edit-based approach inspired by human workflows. PPTAgent first analyzes reference presentations to extract slide-level functional types and content schemas, then drafts an outline and iteratively generates editing actions based on selected reference slides to create new slides. To comprehensively evaluate the quality of generated presentations, we further introduce PPTEval, an evaluation framework that assesses presentations across three dimensions: Content, Design, and Coherence. Results demonstrate that PPTAgent significantly outperforms existing automatic presentation generation methods across all three dimensions.
2501.03937
A precise asymptotic analysis of learning diffusion models: theory and insights
cs.LG cond-mat.dis-nn
In this manuscript, we consider the problem of learning a flow or diffusion-based generative model parametrized by a two-layer auto-encoder, trained with online stochastic gradient descent, on a high-dimensional target density with an underlying low-dimensional manifold structure. We derive a tight asymptotic characterization of low-dimensional projections of the distribution of samples generated by the learned model, ascertaining in particular its dependence on the number of training samples. Building on this analysis, we discuss how mode collapse can arise, and lead to model collapse when the generative model is re-trained on generated synthetic data.
2501.03939
Visual question answering: from early developments to recent advances -- a survey
cs.CV cs.MM
Visual Question Answering (VQA) is an evolving research field aimed at enabling machines to answer questions about visual content by integrating image and language processing techniques such as feature extraction, object detection, text embedding, natural language understanding, and language generation. With the growth of multimodal data research, VQA has gained significant attention due to its broad applications, including interactive educational tools, medical image diagnosis, customer service, entertainment, and social media captioning. Additionally, VQA plays a vital role in assisting visually impaired individuals by generating descriptive content from images. This survey introduces a taxonomy of VQA architectures, categorizing them based on design choices and key components to facilitate comparative analysis and evaluation. We review major VQA approaches, focusing on deep learning-based methods, and explore the emerging field of Large Visual Language Models (LVLMs) that have demonstrated success in multimodal tasks like VQA. The paper further examines available datasets and evaluation metrics essential for measuring VQA system performance, followed by an exploration of real-world VQA applications. Finally, we highlight ongoing challenges and future directions in VQA research, presenting open questions and potential areas for further development. This survey serves as a comprehensive resource for researchers and practitioners interested in the latest advancements and future
2501.03940
Not all tokens are created equal: Perplexity Attention Weighted Networks for AI generated text detection
cs.CL cs.AI
The rapid advancement in large language models (LLMs) has significantly enhanced their ability to generate coherent and contextually relevant text, raising concerns about the misuse of AI-generated content and making it critical to detect it. However, the task remains challenging, particularly in unseen domains or with unfamiliar LLMs. Leveraging LLM next-token distribution outputs offers a theoretically appealing approach for detection, as they encapsulate insights from the models' extensive pre-training on diverse corpora. Despite its promise, zero-shot methods that attempt to operationalize these outputs have met with limited success. We hypothesize that one of the problems is that they use the mean to aggregate next-token distribution metrics across tokens, when some tokens are naturally easier or harder to predict and should be weighted differently. Based on this idea, we propose the Perplexity Attention Weighted Network (PAWN), which uses the last hidden states of the LLM and positions to weight the sum of a series of features based on metrics from the next-token distribution across the sequence length. Although not zero-shot, our method allows us to cache the last hidden states and next-token distribution metrics on disk, greatly reducing the training resource requirements. PAWN shows competitive and even better performance in-distribution than the strongest baselines (fine-tuned LMs) with a fraction of their trainable parameters. Our model also generalizes better to unseen domains and source models, with smaller variability in the decision boundary across distribution shifts. It is also more robust to adversarial attacks, and if the backbone has multilingual capabilities, it presents decent generalization to languages not seen during supervised training, with LLaMA3-1B reaching a mean macro-averaged F1 score of 81.46% in cross-validation with nine languages.
2501.03941
Synthetic Data Privacy Metrics
cs.LG cs.AI
Recent advancements in generative AI have made it possible to create synthetic datasets that can be as accurate as real-world data for training AI models, powering statistical insights, and fostering collaboration with sensitive datasets while offering strong privacy guarantees. Effectively measuring the empirical privacy of synthetic data is an important step in the process. However, while there is a multitude of new privacy metrics being published every day, there currently is no standardization. In this paper, we review the pros and cons of popular metrics that include simulations of adversarial attacks. We also review current best practices for amending generative models to enhance the privacy of the data they create (e.g. differential privacy).
2501.03944
A GPU Implementation of Multi-Guiding Spark Fireworks Algorithm for Efficient Black-Box Neural Network Optimization
cs.NE
Swarm intelligence optimization algorithms have gained significant attention due to their ability to solve complex optimization problems. However, the efficiency of optimization in large-scale problems limits the use of related methods. This paper presents a GPU-accelerated version of the Multi-Guiding Spark Fireworks Algorithm (MGFWA), which significantly improves the computational efficiency compared to its traditional CPU-based counterpart. We benchmark the GPU-MGFWA on several neural network black-box optimization problems and demonstrate its superior performance in terms of both speed and solution quality. By leveraging the parallel processing power of modern GPUs, the proposed GPU-MGFWA results in faster convergence and reduced computation time for large-scale optimization tasks. The proposed implementation offers a promising approach to accelerate swarm intelligence algorithms, making them more suitable for real-time applications and large-scale industrial problems. Source code is released at https://github.com/mxxxr/MGFWA.
2501.03952
Localizing AI: Evaluating Open-Weight Language Models for Languages of Baltic States
cs.CL cs.AI
Although large language models (LLMs) have transformed our expectations of modern language technologies, concerns over data privacy often restrict the use of commercially available LLMs hosted outside of EU jurisdictions. This limits their application in governmental, defence, and other data-sensitive sectors. In this work, we evaluate the extent to which locally deployable open-weight LLMs support lesser-spoken languages such as Lithuanian, Latvian, and Estonian. We examine various size and precision variants of the top-performing multilingual open-weight models, Llama~3, Gemma~2, Phi, and NeMo, on machine translation, multiple-choice question answering, and free-form text generation. The results indicate that while certain models like Gemma~2 perform close to the top commercially available models, many LLMs struggle with these languages. Most surprisingly, however, we find that these models, while showing close to state-of-the-art translation performance, are still prone to lexical hallucinations with errors in at least 1 in 20 words for all open-weight multilingual LLMs.
2501.03957
Vision Language Models as Values Detectors
cs.HC cs.CV
Large Language Models integrating textual and visual inputs have introduced new possibilities for interpreting complex data. Despite their remarkable ability to generate coherent and contextually relevant text based on visual stimuli, the alignment of these models with human perception in identifying relevant elements in images requires further exploration. This paper investigates the alignment between state-of-the-art LLMs and human annotators in detecting elements of relevance within home environment scenarios. We created a set of twelve images depicting various domestic scenarios and enlisted fourteen annotators to identify the key element in each image. We then compared these human responses with outputs from five different LLMs, including GPT-4o and four LLaVA variants. Our findings reveal a varied degree of alignment, with LLaVA 34B showing the highest performance but still scoring low. However, an analysis of the results highlights the models' potential to detect value-laden elements in images, suggesting that with improved training and refined prompts, LLMs could enhance applications in social robotics, assistive technologies, and human-computer interaction by providing deeper insights and more contextually relevant responses.
2501.03961
Channel Coding based on Skew Polynomials and Multivariate Polynomials
cs.IT eess.SP math.IT
This dissertation considers new constructions and decoding approaches for error-correcting codes based on non-conventional polynomials, with the objective of providing new coding solutions to the applications mentioned above. With skew polynomials, we construct codes that are dual-containing, which is a desired property of quantum error-correcting codes. By considering evaluation codes based on skew polynomials, a condition on the existence of optimal support-constrained codes is derived and an application of such codes in the distributed multi-source networks is proposed. For a class of multicast networks, the advantage of vector network coding compared to scalar network coding is investigated. Multivariate polynomials have been attracting increasing interest in constructing codes with repair capabilities by accessing only a small amount of available symbols, which is required to build failure-resistant distributed storage systems. A new class of bivariate evaluation codes and their local recovery capability are studied. Interestingly, the well-known Reed-Solomon codes are used in a class of locally recoverable codes with availability (multiple disjoint recovery sets) via subspace design. Aside from new constructions, decoding approaches are considered in order to increase the error correction capability in the case where the code is fixed. In particular, new lower and upper bounds on the success probability of joint decoding interleaved alternant codes by a syndrome-based decoder are derived, where alternant codes are an important class of algebraic codes containing Goppa codes, BCH codes, and Reed-Muller codes as sub-classes.
2501.03964
A comparative study of uncertainty quantification methods in gust response analysis of a Lift-Plus-Cruise eVTOL aircraft wing
cs.CE
Wind gusts, being inherently stochastic, can significantly influence the safety and performance of aircraft. This study investigates a three-dimensional uncertainty quantification (UQ) problem to explore how uncertainties in gust and flight conditions affect the structural response of a Lift-Plus-Cruise eVTOL aircraft wing. The analysis employs an unsteady aeroelastic model with a one-way coupling between a panel method aerodynamic solver and a shell analysis structural solver to predict the wing's response under varying conditions. Additionally, this paper presents a comparative evaluation of commonly used non-intrusive UQ methods, including non-intrusive polynomial chaos, kriging, Monte Carlo, univariate dimension reduction, and gradient-enhanced univariate dimension reduction. These methods are assessed based on their effectiveness in estimating various risk measures-mean, standard deviation, and 95th percentile-of critical structural response outputs such as maximum tip displacement and average strain energy. The numerical results reveal significant variability in the structural response outputs, even under relatively small ranges of uncertain inputs. This highlights the sensitivity of the system to uncertainties in gust and flight conditions. Furthermore, the performance of the implemented UQ methods varies significantly depending on the specific risk measures and the quantity of interest being analyzed.
2501.03967
Temporal Feature Weaving for Neonatal Echocardiographic Viewpoint Video Classification
cs.CV
Automated viewpoint classification in echocardiograms can help under-resourced clinics and hospitals in providing faster diagnosis and screening when expert technicians may not be available. We propose a novel approach towards echocardiographic viewpoint classification. We show that treating viewpoint classification as video classification rather than image classification yields advantage. We propose a CNN-GRU architecture with a novel temporal feature weaving method, which leverages both spatial and temporal information to yield a 4.33\% increase in accuracy over baseline image classification while using only four consecutive frames. The proposed approach incurs minimal computational overhead. Additionally, we publish the Neonatal Echocardiogram Dataset (NED), a professionally-annotated dataset providing sixteen viewpoints and associated echocardipgraphy videos to encourage future work and development in this field. Code available at: https://github.com/satchelfrench/NED
2501.03968
VLM-driven Behavior Tree for Context-aware Task Planning
cs.RO cs.AI cs.CV cs.HC
The use of Large Language Models (LLMs) for generating Behavior Trees (BTs) has recently gained attention in the robotics community, yet remains in its early stages of development. In this paper, we propose a novel framework that leverages Vision-Language Models (VLMs) to interactively generate and edit BTs that address visual conditions, enabling context-aware robot operations in visually complex environments. A key feature of our approach lies in the conditional control through self-prompted visual conditions. Specifically, the VLM generates BTs with visual condition nodes, where conditions are expressed as free-form text. Another VLM process integrates the text into its prompt and evaluates the conditions against real-world images during robot execution. We validated our framework in a real-world cafe scenario, demonstrating both its feasibility and limitations.
2501.03971
Impact of Leg Stiffness on Energy Efficiency in One Legged Hopping
cs.RO
In the fields of robotics and biomechanics, the integration of elastic elements such as springs and tendons in legged systems has long been recognized for enabling energy-efficient locomotion. Yet, a significant challenge persists: designing a robotic leg that perform consistently across diverse operating conditions, especially varying average forward speeds. It remains unclear whether, for such a range of operating conditions, the stiffness of the elastic elements needs to be varied or if a similar performance can be obtained by changing the motion and actuation while keeping the stiffness fixed. This work explores the influence of the leg stiffness on the energy efficiency of a monopedal robot through an extensive parametric study of its periodic hopping motion. To this end, we formulate an optimal control problem parameterized by average forward speed and leg stiffness, solving it numerically using direct collocation. Our findings indicate that, compared to the use of a fixed stiffness, employing variable stiffness in legged systems improves energy efficiency by 20 % maximally and by 6.8 % on average across a range of speeds.
2501.03972
MAD-BA: 3D LiDAR Bundle Adjustment -- from Uncertainty Modelling to Structure Optimization
cs.RO
The joint optimization of sensor poses and 3D structure is fundamental for state estimation in robotics and related fields. Current LiDAR systems often prioritize pose optimization, with structure refinement either omitted or treated separately using representations like signed distance functions or neural networks. This paper introduces a framework for simultaneous optimization of sensor poses and 3D map, represented as surfels. A generalized LiDAR uncertainty model is proposed to address degraded or less reliable measurements in varying scenarios. Experimental results on public datasets demonstrate improved performance over most comparable state-of-the-art methods. The system is provided as open-source software to support further research.
2501.03988
Semantically Cohesive Word Grouping in Indian Languages
cs.CL
Indian languages are inflectional and agglutinative and typically follow clause-free word order. The structure of sentences across most major Indian languages are similar when their dependency parse trees are considered. While some differences in the parsing structure occur due to peculiarities of a language or its preferred natural way of conveying meaning, several apparent differences are simply due to the granularity of representation of the smallest semantic unit of processing in a sentence. The semantic unit is typically a word, typographically separated by whitespaces. A single whitespace-separated word in one language may correspond to a group of words in another. Hence, grouping of words based on semantics helps unify the parsing structure of parallel sentences across languages and, in the process, morphology. In this work, we propose word grouping as a major preprocessing step for any computational or linguistic processing of sentences for Indian languages. Among Indian languages, since Hindi is one of the least agglutinative, we expect it to benefit the most from word-grouping. Hence, in this paper, we focus on Hindi to study the effects of grouping. We perform quantitative assessment of our proposal with an intrinsic method that perturbs sentences by shuffling words as well as an extrinsic evaluation that verifies the importance of word grouping for the task of Machine Translation (MT) using decomposed prompting. We also qualitatively analyze certain aspects of the syntactic structure of sentences. Our experiments and analyses show that the proposed grouping technique brings uniformity in the syntactic structures, as well as aids underlying NLP tasks.
2501.03989
(De)-Indexing and the Right to be Forgotten
cs.CY cs.IR
In the digital age, the challenge of forgetfulness has emerged as a significant concern, particularly regarding the management of personal data and its accessibility online. The right to be forgotten (RTBF) allows individuals to request the removal of outdated or harmful information from public access, yet implementing this right poses substantial technical difficulties for search engines. This paper aims to introduce non-experts to the foundational concepts of information retrieval (IR) and de-indexing, which are critical for understanding how search engines can effectively "forget" certain content. We will explore various IR models, including boolean, probabilistic, vector space, and embedding-based approaches, as well as the role of Large Language Models (LLMs) in enhancing data processing capabilities. By providing this overview, we seek to highlight the complexities involved in balancing individual privacy rights with the operational challenges faced by search engines in managing information visibility.
2501.03991
Influences on LLM Calibration: A Study of Response Agreement, Loss Functions, and Prompt Styles
cs.CL
Calibration, the alignment between model confidence and prediction accuracy, is critical for the reliable deployment of large language models (LLMs). Existing works neglect to measure the generalization of their methods to other prompt styles and different sizes of LLMs. To address this, we define a controlled experimental setting covering 12 LLMs and four prompt styles. We additionally investigate if incorporating the response agreement of multiple LLMs and an appropriate loss function can improve calibration performance. Concretely, we build Calib-n, a novel framework that trains an auxiliary model for confidence estimation that aggregates responses from multiple LLMs to capture inter-model agreement. To optimize calibration, we integrate focal and AUC surrogate losses alongside binary cross-entropy. Experiments across four datasets demonstrate that both response agreement and focal loss improve calibration from baselines. We find that few-shot prompts are the most effective for auxiliary model-based methods, and auxiliary models demonstrate robust calibration performance across accuracy variations, outperforming LLMs' internal probabilities and verbalized confidences. These insights deepen the understanding of influence factors in LLM calibration, supporting their reliable deployment in diverse applications.
2501.03992
NeuralSVG: An Implicit Representation for Text-to-Vector Generation
cs.CV
Vector graphics are essential in design, providing artists with a versatile medium for creating resolution-independent and highly editable visual content. Recent advancements in vision-language and diffusion models have fueled interest in text-to-vector graphics generation. However, existing approaches often suffer from over-parameterized outputs or treat the layered structure - a core feature of vector graphics - as a secondary goal, diminishing their practical use. Recognizing the importance of layered SVG representations, we propose NeuralSVG, an implicit neural representation for generating vector graphics from text prompts. Inspired by Neural Radiance Fields (NeRFs), NeuralSVG encodes the entire scene into the weights of a small MLP network, optimized using Score Distillation Sampling (SDS). To encourage a layered structure in the generated SVG, we introduce a dropout-based regularization technique that strengthens the standalone meaning of each shape. We additionally demonstrate that utilizing a neural representation provides an added benefit of inference-time control, enabling users to dynamically adapt the generated SVG based on user-provided inputs, all with a single learned representation. Through extensive qualitative and quantitative evaluations, we demonstrate that NeuralSVG outperforms existing methods in generating structured and flexible SVG.
2501.03995
RAG-Check: Evaluating Multimodal Retrieval Augmented Generation Performance
cs.LG cs.CV cs.IR cs.IT math.IT
Retrieval-augmented generation (RAG) improves large language models (LLMs) by using external knowledge to guide response generation, reducing hallucinations. However, RAG, particularly multi-modal RAG, can introduce new hallucination sources: (i) the retrieval process may select irrelevant pieces (e.g., documents, images) as raw context from the database, and (ii) retrieved images are processed into text-based context via vision-language models (VLMs) or directly used by multi-modal language models (MLLMs) like GPT-4o, which may hallucinate. To address this, we propose a novel framework to evaluate the reliability of multi-modal RAG using two performance measures: (i) the relevancy score (RS), assessing the relevance of retrieved entries to the query, and (ii) the correctness score (CS), evaluating the accuracy of the generated response. We train RS and CS models using a ChatGPT-derived database and human evaluator samples. Results show that both models achieve ~88% accuracy on test data. Additionally, we construct a 5000-sample human-annotated database evaluating the relevancy of retrieved pieces and the correctness of response statements. Our RS model aligns with human preferences 20% more often than CLIP in retrieval, and our CS model matches human preferences ~91% of the time. Finally, we assess various RAG systems' selection and generation performances using RS and CS.
2501.03999
WAPTS: A Weighted Allocation Probability Adjusted Thompson Sampling Algorithm for High-Dimensional and Sparse Experiment Settings
cs.LG stat.ML
Aiming for more effective experiment design, such as in video content advertising where different content options compete for user engagement, these scenarios can be modeled as multi-arm bandit problems. In cases where limited interactions are available due to external factors, such as the cost of conducting experiments, recommenders often face constraints due to the small number of user interactions. In addition, there is a trade-off between selecting the best treatment and the ability to personalize and contextualize based on individual factors. A popular solution to this dilemma is the Contextual Bandit framework. It aims to maximize outcomes while incorporating personalization (contextual) factors, customizing treatments such as a user's profile to individual preferences. Despite their advantages, Contextual Bandit algorithms face challenges like measurement bias and the 'curse of dimensionality.' These issues complicate the management of numerous interventions and often lead to data sparsity through participant segmentation. To address these problems, we introduce the Weighted Allocation Probability Adjusted Thompson Sampling (WAPTS) algorithm. WAPTS builds on the contextual Thompson Sampling method by using a dynamic weighting parameter. This improves the allocation process for interventions and enables rapid optimization in data-sparse environments. We demonstrate the performance of our approach on different numbers of arms and effect sizes.
2501.04000
A Survey on Federated Learning in Human Sensing
cs.LG cs.HC
Human Sensing, a field that leverages technology to monitor human activities, psycho-physiological states, and interactions with the environment, enhances our understanding of human behavior and drives the development of advanced services that improve overall quality of life. However, its reliance on detailed and often privacy-sensitive data as the basis for its machine learning (ML) models raises significant legal and ethical concerns. The recently proposed ML approach of Federated Learning (FL) promises to alleviate many of these concerns, as it is able to create accurate ML models without sending raw user data to a central server. While FL has demonstrated its usefulness across a variety of areas, such as text prediction and cyber security, its benefits in Human Sensing are under-explored, given the particular challenges in this domain. This survey conducts a comprehensive analysis of the current state-of-the-art studies on FL in Human Sensing, and proposes a taxonomy and an eight-dimensional assessment for FL approaches. Through the eight-dimensional assessment, we then evaluate whether the surveyed studies consider a specific FL-in-Human-Sensing challenge or not. Finally, based on the overall analysis, we discuss open challenges and highlight five research aspects related to FL in Human Sensing that require urgent research attention. Our work provides a comprehensive corpus of FL studies and aims to assist FL practitioners in developing and evaluating solutions that effectively address the real-world complexities of Human Sensing.
2501.04001
Sa2VA: Marrying SAM2 with LLaVA for Dense Grounded Understanding of Images and Videos
cs.CV
This work presents Sa2VA, the first unified model for dense grounded understanding of both images and videos. Unlike existing multi-modal large language models, which are often limited to specific modalities and tasks, Sa2VA supports a wide range of image and video tasks, including referring segmentation and conversation, with minimal one-shot instruction tuning. Sa2VA combines SAM-2, a foundation video segmentation model, with LLaVA, an advanced vision-language model, and unifies text, image, and video into a shared LLM token space. Using the LLM, Sa2VA generates instruction tokens that guide SAM-2 in producing precise masks, enabling a grounded, multi-modal understanding of both static and dynamic visual content. Additionally, we introduce Ref-SAV, an auto-labeled dataset containing over 72k object expressions in complex video scenes, designed to boost model performance. We also manually validate 2k video objects in the Ref-SAV datasets to benchmark referring video object segmentation in complex environments. Experiments show that Sa2VA achieves state-of-the-art across multiple tasks, particularly in referring video object segmentation, highlighting its potential for complex real-world applications.
2501.04002
Extraction Of Cumulative Blobs From Dynamic Gestures
cs.CV
Gesture recognition is a perceptual user interface, which is based on CV technology that allows the computer to interpret human motions as commands, allowing users to communicate with a computer without the use of hands, thus making the mouse and keyboard superfluous. Gesture recognition's main weakness is a light condition because gesture control is based on computer vision, which heavily relies on cameras. These cameras are used to interpret gestures in 2D and 3D, so the extracted information can vary depending on the source of light. The limitation of the system cannot work in a dark environment. A simple night vision camera can be used as our camera for motion capture as they also blast out infrared light which is not visible to humans but can be clearly seen with a camera that has no infrared filter this majorly overcomes the limitation of systems which cannot work in a dark environment. So, the video stream from the camera is fed into a Raspberry Pi which has a Python program running OpenCV module which is used for detecting, isolating and tracking the path of dynamic gesture, then we use an algorithm of machine learning to recognize the pattern drawn and accordingly control the GPIOs of the raspberry pi to perform some activities.
2501.04003
Are VLMs Ready for Autonomous Driving? An Empirical Study from the Reliability, Data, and Metric Perspectives
cs.CV cs.RO
Recent advancements in Vision-Language Models (VLMs) have sparked interest in their use for autonomous driving, particularly in generating interpretable driving decisions through natural language. However, the assumption that VLMs inherently provide visually grounded, reliable, and interpretable explanations for driving remains largely unexamined. To address this gap, we introduce DriveBench, a benchmark dataset designed to evaluate VLM reliability across 17 settings (clean, corrupted, and text-only inputs), encompassing 19,200 frames, 20,498 question-answer pairs, three question types, four mainstream driving tasks, and a total of 12 popular VLMs. Our findings reveal that VLMs often generate plausible responses derived from general knowledge or textual cues rather than true visual grounding, especially under degraded or missing visual inputs. This behavior, concealed by dataset imbalances and insufficient evaluation metrics, poses significant risks in safety-critical scenarios like autonomous driving. We further observe that VLMs struggle with multi-modal reasoning and display heightened sensitivity to input corruptions, leading to inconsistencies in performance. To address these challenges, we propose refined evaluation metrics that prioritize robust visual grounding and multi-modal understanding. Additionally, we highlight the potential of leveraging VLMs' awareness of corruptions to enhance their reliability, offering a roadmap for developing more trustworthy and interpretable decision-making systems in real-world autonomous driving contexts. The benchmark toolkit is publicly accessible.
2501.04004
LiMoE: Mixture of LiDAR Representation Learners from Automotive Scenes
cs.CV cs.LG cs.RO
LiDAR data pretraining offers a promising approach to leveraging large-scale, readily available datasets for enhanced data utilization. However, existing methods predominantly focus on sparse voxel representation, overlooking the complementary attributes provided by other LiDAR representations. In this work, we propose LiMoE, a framework that integrates the Mixture of Experts (MoE) paradigm into LiDAR data representation learning to synergistically combine multiple representations, such as range images, sparse voxels, and raw points. Our approach consists of three stages: i) Image-to-LiDAR Pretraining, which transfers prior knowledge from images to point clouds across different representations; ii) Contrastive Mixture Learning (CML), which uses MoE to adaptively activate relevant attributes from each representation and distills these mixed features into a unified 3D network; iii) Semantic Mixture Supervision (SMS), which combines semantic logits from multiple representations to boost downstream segmentation performance. Extensive experiments across 11 large-scale LiDAR datasets demonstrate our effectiveness and superiority. The code and model checkpoints have been made publicly accessible.
2501.04005
LargeAD: Large-Scale Cross-Sensor Data Pretraining for Autonomous Driving
cs.CV cs.LG cs.RO
Recent advancements in vision foundation models (VFMs) have revolutionized visual perception in 2D, yet their potential for 3D scene understanding, particularly in autonomous driving applications, remains underexplored. In this paper, we introduce LargeAD, a versatile and scalable framework designed for large-scale 3D pretraining across diverse real-world driving datasets. Our framework leverages VFMs to extract semantically rich superpixels from 2D images, which are aligned with LiDAR point clouds to generate high-quality contrastive samples. This alignment facilitates cross-modal representation learning, enhancing the semantic consistency between 2D and 3D data. We introduce several key innovations: i) VFM-driven superpixel generation for detailed semantic representation, ii) a VFM-assisted contrastive learning strategy to align multimodal features, iii) superpoint temporal consistency to maintain stable representations across time, and iv) multi-source data pretraining to generalize across various LiDAR configurations. Our approach delivers significant performance improvements over state-of-the-art methods in both linear probing and fine-tuning tasks for both LiDAR-based segmentation and object detection. Extensive experiments on eleven large-scale multi-modal datasets highlight our superior performance, demonstrating the adaptability, efficiency, and robustness in real-world autonomous driving scenarios.
2501.04006
Advancing Similarity Search with GenAI: A Retrieval Augmented Generation Approach
cs.IR
This article introduces an innovative Retrieval Augmented Generation approach to similarity search. The proposed method uses a generative model to capture nuanced semantic information and retrieve similarity scores based on advanced context understanding. The study focuses on the BIOSSES dataset containing 100 pairs of sentences extracted from the biomedical domain, and introduces similarity search correlation results that outperform those previously attained on this dataset. Through an in-depth analysis of the model sensitivity, the research identifies optimal conditions leading to the highest similarity search accuracy: the results reveals high Pearson correlation scores, reaching specifically 0.905 at a temperature of 0.5 and a sample size of 20 examples provided in the prompt. The findings underscore the potential of generative models for semantic information retrieval and emphasize a promising research direction to similarity search.
2501.04007
Untapped Potential in Self-Optimization of Hopfield Networks: The Creativity of Unsupervised Learning
cs.NE nlin.AO
The Self-Optimization (SO) model can be considered as the third operational mode of the classical Hopfield Network (HN), leveraging the power of associative memory to enhance optimization performance. Moreover, is has been argued to express characteristics of minimal agency which, together with its biological plausibility, renders it useful for the study of artificial life. In this article, we draw attention to another facet of the SO model: its capacity for creativity. Drawing on the creativity studies literature, we argue that the model satisfies the necessary and sufficient conditions of a creative process. Moreover, we explore the dependency of different creative outcomes based on learning parameters, specifically the learning and reset rates. We conclude that the SO model allows for simulating and understanding the emergence of creative behaviors in artificial systems that learn.
2501.04008
A Generative AI-driven Metadata Modelling Approach
cs.DL cs.AI cs.IR
Since decades, the modelling of metadata has been core to the functioning of any academic library. Its importance has only enhanced with the increasing pervasiveness of Generative Artificial Intelligence (AI)-driven information activities and services which constitute a library's outreach. However, with the rising importance of metadata, there arose several outstanding problems with the process of designing a library metadata model impacting its reusability, crosswalk and interoperability with other metadata models. This paper posits that the above problems stem from an underlying thesis that there should only be a few core metadata models which would be necessary and sufficient for any information service using them, irrespective of the heterogeneity of intra-domain or inter-domain settings. To that end, this paper advances a contrary view of the above thesis and substantiates its argument in three key steps. First, it introduces a novel way of thinking about a library metadata model as an ontology-driven composition of five functionally interlinked representation levels from perception to its intensional definition via properties. Second, it introduces the representational manifoldness implicit in each of the five levels which cumulatively contributes to a conceptually entangled library metadata model. Finally, and most importantly, it proposes a Generative AI-driven Human-Large Language Model (LLM) collaboration based metadata modelling approach to disentangle the entanglement inherent in each representation level leading to the generation of a conceptually disentangled metadata model. Throughout the paper, the arguments are exemplified by motivating scenarios and examples from representative libraries handling cancer information.
2501.04009
Multi-SpaCE: Multi-Objective Subsequence-based Sparse Counterfactual Explanations for Multivariate Time Series Classification
cs.NE cs.LG stat.ML
Deep Learning systems excel in complex tasks but often lack transparency, limiting their use in critical applications. Counterfactual explanations, a core tool within eXplainable Artificial Intelligence (XAI), offer insights into model decisions by identifying minimal changes to an input to alter its predicted outcome. However, existing methods for time series data are limited by univariate assumptions, rigid constraints on modifications, or lack of validity guarantees. This paper introduces Multi-SpaCE, a multi-objective counterfactual explanation method for multivariate time series. Using non-dominated ranking genetic algorithm II (NSGA-II), Multi-SpaCE balances proximity, sparsity, plausibility, and contiguity. Unlike most methods, it ensures perfect validity, supports multivariate data and provides a Pareto front of solutions, enabling flexibility to different end-user needs. Comprehensive experiments in diverse datasets demonstrate the ability of Multi-SpaCE to consistently achieve perfect validity and deliver superior performance compared to existing methods.
2501.04012
FlexCache: Flexible Approximate Cache System for Video Diffusion
cs.MM cs.LG
Text-to-Video applications receive increasing attention from the public. Among these, diffusion models have emerged as the most prominent approach, offering impressive quality in visual content generation. However, it still suffers from substantial computational complexity, often requiring several minutes to generate a single video. While prior research has addressed the computational overhead in text-to-image diffusion models, the techniques developed are not directly suitable for video diffusion models due to the significantly larger cache requirements and enhanced computational demands associated with video generation. We present FlexCache, a flexible approximate cache system that addresses the challenges in two main designs. First, we compress the caches before saving them to storage. Our compression strategy can reduce 6.7 times consumption on average. Then we find that the approximate cache system can achieve higher hit rate and computation savings by decoupling the object and background. We further design a tailored cache replacement policy to support the two techniques mentioned above better. Through our evaluation, FlexCache reaches 1.26 times higher throughput and 25% lower cost compared to the state-of-the-art diffusion approximate cache system.
2501.04014
AICat: An AI Cataloguing Approach to Support the EU AI Act
cs.DL cs.AI cs.CY
The European Union's Artificial Intelligence Act (AI Act) requires providers and deployers of high-risk AI applications to register their systems into the EU database, wherein the information should be represented and maintained in an easily-navigable and machine-readable manner. Given the uptake of open data and Semantic Web-based approaches for other EU repositories, in particular the use of the Data Catalogue vocabulary Application Profile (DCAT-AP), a similar solution for managing the EU database of high-risk AI systems is needed. This paper introduces AICat - an extension of DCAT for representing catalogues of AI systems that provides consistency, machine-readability, searchability, and interoperability in managing open metadata regarding AI systems. This open approach to cataloguing ensures transparency, traceability, and accountability in AI application markets beyond the immediate needs of high-risk AI compliance in the EU. AICat is available online at https://w3id.org/aicat under the CC-BY-4.0 license.
2501.04018
MERCURY: A fast and versatile multi-resolution based global emulator of compound climate hazards
physics.ao-ph cs.LG stat.AP
High-impact climate damages are often driven by compounding climate conditions. For example, elevated heat stress conditions can arise from a combination of high humidity and temperature. To explore future changes in compounding hazards under a range of climate scenarios and with large ensembles, climate emulators can provide light-weight, data-driven complements to Earth System Models. Yet, only a few existing emulators can jointly emulate multiple climate variables. In this study, we present the Multi-resolution EmulatoR for CompoUnd climate Risk analYsis: MERCURY. MERCURY extends multi-resolution analysis to a spatio-temporal framework for versatile emulation of multiple variables. MERCURY leverages data-driven, image compression techniques to generate emulations in a memory-efficient manner. MERCURY consists of a regional component that represents the monthly, regional response of a given variable to yearly Global Mean Temperature (GMT) using a probabilistic regression based additive model, resolving regional cross-correlations. It then adapts a reverse lifting-scheme operator to jointly spatially disaggregate regional, monthly values to grid-cell level. We demonstrate MERCURY's capabilities on representing the humid-heat metric, Wet Bulb Globe Temperature, as derived from temperature and relative humidity emulations. The emulated WBGT spatial correlations correspond well to those of ESMs and the 95% and 97.5% quantiles of WBGT distributions are well captured, with an average of 5% deviation. MERCURY's setup allows for region-specific emulations from which one can efficiently "zoom" into the grid-cell level across multiple variables by means of the reverse lifting-scheme operator. This circumvents the traditional problem of having to emulate complete, global-fields of climate data and resulting storage requirements.
2501.04023
Approximation Rates in Fr\'echet Metrics: Barron Spaces, Paley-Wiener Spaces, and Fourier Multipliers
math.NA cs.IT cs.LG cs.NA math.IT stat.ML
Operator learning is a recent development in the simulation of Partial Differential Equations (PDEs) by means of neural networks. The idea behind this approach is to learn the behavior of an operator, such that the resulting neural network is an (approximate) mapping in infinite-dimensional spaces that is capable of (approximately) simulating the solution operator governed by the PDE. In our work, we study some general approximation capabilities for linear differential operators by approximating the corresponding symbol in the Fourier domain. Analogous to the structure of the class of H\"ormander-Symbols, we consider the approximation with respect to a topology that is induced by a sequence of semi-norms. In that sense, we measure the approximation error in terms of a Fr\'echet metric, and our main result identifies sufficient conditions for achieving a predefined approximation error. Secondly, we then focus on a natural extension of our main theorem, in which we manage to reduce the assumptions on the sequence of semi-norms. Based on existing approximation results for the exponential spectral Barron space, we then present a concrete example of symbols that can be approximated well.
2501.04038
Listening and Seeing Again: Generative Error Correction for Audio-Visual Speech Recognition
cs.MM cs.AI cs.SD eess.AS
Unlike traditional Automatic Speech Recognition (ASR), Audio-Visual Speech Recognition (AVSR) takes audio and visual signals simultaneously to infer the transcription. Recent studies have shown that Large Language Models (LLMs) can be effectively used for Generative Error Correction (GER) in ASR by predicting the best transcription from ASR-generated N-best hypotheses. However, these LLMs lack the ability to simultaneously understand audio and visual, making the GER approach challenging to apply in AVSR. In this work, we propose a novel GER paradigm for AVSR, termed AVGER, that follows the concept of ``listening and seeing again''. Specifically, we first use the powerful AVSR system to read the audio and visual signals to get the N-Best hypotheses, and then use the Q-former-based Multimodal Synchronous Encoder to read the audio and visual information again and convert them into an audio and video compression representation respectively that can be understood by LLM. Afterward, the audio-visual compression representation and the N-Best hypothesis together constitute a Cross-modal Prompt to guide the LLM in producing the best transcription. In addition, we also proposed a Multi-Level Consistency Constraint training criterion, including logits-level, utterance-level and representations-level, to improve the correction accuracy while enhancing the interpretability of audio and visual compression representations. The experimental results on the LRS3 dataset show that our method outperforms current mainstream AVSR systems. The proposed AVGER can reduce the Word Error Rate (WER) by 24% compared to them. Code and models can be found at: https://github.com/CircleRedRain/AVGER.
2501.04040
A Survey on Large Language Models with some Insights on their Capabilities and Limitations
cs.CL cs.AI cs.LG cs.NE
The rapid advancement of artificial intelligence, particularly with the development of Large Language Models (LLMs) built on the transformer architecture, has redefined the capabilities of natural language processing. These models now exhibit remarkable performance across various language-related tasks, such as text generation, question answering, translation, and summarization, often rivaling human-like comprehension. More intriguingly, LLMs have demonstrated emergent abilities extending beyond their core functions, showing proficiency in tasks like commonsense reasoning, code generation, and arithmetic. This survey paper explores the foundational components, scaling mechanisms, and architectural strategies that drive these capabilities. Emphasizing models like GPT and LLaMA, we analyze the impact of exponential data and computational growth on LLM performance, while also addressing the trade-offs associated with scaling. We also examine LLM applications across sectors, such as healthcare, finance, education, and law, highlighting their adaptability and potential to solve domain-specific challenges. Central to this work are the questions of how LLMs generalize across diverse tasks, exhibit planning, and reasoning abilities, and whether these emergent abilities can be systematically elicited or enhanced. In particular, we provide some insights into the CoT (Chain of Thought) and PoT (Plan of Thought) abilities within LLMs, focusing on how pre-training data influences their emergence. Additionally, we investigate LLM-modulo frameworks that integrate external systems, allowing LLMs to handle complex, dynamic tasks. By analyzing these factors, this paper aims to foster the ongoing discussion on the capabilities and limits of LLMs, promoting their responsible development and application in novel and increasingly complex environments.
2501.04046
Traits of a Leader: User Influence Level Prediction through Sociolinguistic Modeling
physics.soc-ph cs.AI cs.CY
Recognition of a user's influence level has attracted much attention as human interactions move online. Influential users have the ability to sway others' opinions to achieve some goals. As a result, predicting users' level of influence can help to understand social networks, forecast trends, prevent misinformation, etc. However, predicting user influence is a challenging problem because the concept of influence is specific to a situation or a domain, and user communications are limited to text. In this work, we define user influence level as a function of community endorsement and develop a model that significantly outperforms the baseline by leveraging demographic and personality data. This approach consistently improves RankDCG scores across eight different domains.
2501.04052
The Power of Negative Zero: Datatype Customization for Quantized Large Language Models
cs.LG cs.CL
Large language models (LLMs) have demonstrated remarkable performance across various machine learning tasks, quickly becoming one of the most prevalent AI workloads. Yet the substantial memory requirement of LLMs significantly hinders their deployment for end users. Post-training quantization (PTQ) serves as one of the most hardware-efficient methods to mitigate the memory and computational demands of LLMs. Although the traditional integer (INT) datatype has received widespread adoption in PTQ methods, floating-point (FP) quantization has emerged as a viable alternative thanks to its effectiveness in fitting LLM numerical distributions. However, the FP datatype in sign-magnitude binary representation contains both positive and negative zero, which constrains its representation capability, particularly under low precision (3 and 4 bits). In this paper, we extend the basic FP datatype to perform Redundant Zero Remapping (RaZeR), which remaps the negative zero FP encoding to a set of pre-defined special values to maximally utilize FP quantization encodings and to better fit LLM numerical distributions. Through careful selection of special values, RaZeR outperforms conventional asymmetric INT quantization while achieving high computational efficiency. We demonstrate that RaZeR can be seamlessly integrated with quantization algorithms for both weights and KV-cache, including advanced methods with clipping and transformations, and consistently achieve better model accuracy. Additionally, we implement a fast GEMV kernel with fused dequantization that efficiently converts the 4-bit RaZeR value to FP16 through novel bit-level manipulation. On modern GPUs, our evaluation shows that RaZeR improves the GEMV speed by up to 7.56$\times$ compared to the FP16 implementation, while achieving up to 2.72$\times$ speedup in the LLM decoding throughput.
2501.04060
SFADNet: Spatio-temporal Fused Graph based on Attention Decoupling Network for Traffic Prediction
cs.LG
In recent years, traffic flow prediction has played a crucial role in the management of intelligent transportation systems. However, traditional prediction methods are often limited by static spatial modeling, making it difficult to accurately capture the dynamic and complex relationships between time and space, thereby affecting prediction accuracy. This paper proposes an innovative traffic flow prediction network, SFADNet, which categorizes traffic flow into multiple traffic patterns based on temporal and spatial feature matrices. For each pattern, we construct an independent adaptive spatio-temporal fusion graph based on a cross-attention mechanism, employing residual graph convolution modules and time series modules to better capture dynamic spatio-temporal relationships under different fine-grained traffic patterns. Extensive experimental results demonstrate that SFADNet outperforms current state-of-the-art baselines across four large-scale datasets.
2501.04061
Causal Machine Learning Methods for Estimating Personalised Treatment Effects -- Insights on validity from two large trials
cs.LG stat.ML
Causal machine learning (ML) methods hold great promise for advancing precision medicine by estimating personalized treatment effects. However, their reliability remains largely unvalidated in empirical settings. In this study, we assessed the internal and external validity of 17 mainstream causal heterogeneity ML methods -- including metalearners, tree-based methods, and deep learning methods -- using data from two large randomized controlled trials: the International Stroke Trial (N=19,435) and the Chinese Acute Stroke Trial (N=21,106). Our findings reveal that none of the ML methods reliably validated their performance, neither internal nor external, showing significant discrepancies between training and test data on the proposed evaluation metrics. The individualized treatment effects estimated from training data failed to generalize to the test data, even in the absence of distribution shifts. These results raise concerns about the current applicability of causal ML models in precision medicine, and highlight the need for more robust validation techniques to ensure generalizability.
2501.04062
ChronoLLM: A Framework for Customizing Large Language Model for Digital Twins generalization based on PyChrono
cs.SE cs.AI cs.CE
Recently, the integration of advanced simulation technologies with artificial intelligence (AI) is revolutionizing science and engineering research. ChronoLlama introduces a novel framework that customizes the open-source LLMs, specifically for code generation, paired with PyChrono for multi-physics simulations. This integration aims to automate and improve the creation of simulation scripts, thus enhancing model accuracy and efficiency. This combination harnesses the speed of AI-driven code generation with the reliability of physics-based simulations, providing a powerful tool for researchers and engineers. Empirical results indicate substantial enhancements in simulation setup speed, accuracy of the generated codes, and overall computational efficiency. ChronoLlama not only expedites the development and testing of multibody systems but also spearheads a scalable, AI-enhanced approach to managing intricate mechanical simulations. This pioneering integration of cutting-edge AI with traditional simulation platforms represents a significant leap forward in automating and optimizing design processes in engineering applications.
2501.04063
Fuzzy Information Entropy and Region Biased Matrix Factorization for Web Service QoS Prediction
cs.LG
Nowadays, there are many similar services available on the internet, making Quality of Service (QoS) a key concern for users. Since collecting QoS values for all services through user invocations is impractical, predicting QoS values is a more feasible approach. Matrix factorization is considered an effective prediction method. However, most existing matrix factorization algorithms focus on capturing global similarities between users and services, overlooking the local similarities between users and their similar neighbors, as well as the non-interactive effects between users and services. This paper proposes a matrix factorization approach based on user information entropy and region bias, which utilizes a similarity measurement method based on fuzzy information entropy to identify similar neighbors of users. Simultaneously, it integrates the region bias between each user and service linearly into matrix factorization to capture the non-interactive features between users and services. This method demonstrates improved predictive performance in more realistic and complex network environments. Additionally, numerous experiments are conducted on real-world QoS datasets. The experimental results show that the proposed method outperforms some of the state-of-the-art methods in the field at matrix densities ranging from 5% to 20%.
2501.04066
FedKD-hybrid: Federated Hybrid Knowledge Distillation for Lithography Hotspot Detection
cs.LG cs.AR
Federated Learning (FL) provides novel solutions for machine learning (ML)-based lithography hotspot detection (LHD) under distributed privacy-preserving settings. Currently, two research pipelines have been investigated to aggregate local models and achieve global consensus, including parameter/nonparameter based (also known as knowledge distillation, namely KD). While these two kinds of methods show effectiveness in specific scenarios, we note they have not fully utilized and transferred the information learned, leaving the potential of FL-based LDH remains unexplored. Thus, we propose FedKDhybrid in this study to mitigate the research gap. Specifically, FedKD-hybrid clients agree on several identical layers across all participants and a public dataset for achieving global consensus. During training, the trained local model will be evaluated on the public dataset, and the generated logits will be uploaded along with the identical layer parameters. The aggregated information is consequently used to update local models via the public dataset as a medium. We compare our proposed FedKD-hybrid with several state-of-the-art (SOTA) FL methods under ICCAD-2012 and FAB (real-world collected) datasets with different settings; the experimental results demonstrate the superior performance of the FedKD-hybrid algorithm. Our code is available at https://github.com/itsnotacie/NN-FedKD-hybrid
2501.04067
Explainable Time Series Prediction of Tyre Energy in Formula One Race Strategy
cs.LG cs.AI
Formula One (F1) race strategy takes place in a high-pressure and fast-paced environment where split-second decisions can drastically affect race results. Two of the core decisions of race strategy are when to make pit stops (i.e. replace the cars' tyres) and which tyre compounds (hard, medium or soft, in normal conditions) to select. The optimal pit stop decisions can be determined by estimating the tyre degradation of these compounds, which in turn can be computed from the energy applied to each tyre, i.e. the tyre energy. In this work, we trained deep learning models, using the Mercedes-AMG PETRONAS F1 team's historic race data consisting of telemetry, to forecast tyre energies during races. Additionally, we fitted XGBoost, a decision tree-based machine learning algorithm, to the same dataset and compared the results, with both giving impressive performance. Furthermore, we incorporated two different explainable AI methods, namely feature importance and counterfactual explanations, to gain insights into the reasoning behind the forecasts. Our contributions thus result in an explainable, automated method which could assist F1 teams in optimising their race strategy.
2501.04068
Explainable Reinforcement Learning for Formula One Race Strategy
cs.LG cs.AI
In Formula One, teams compete to develop their cars and achieve the highest possible finishing position in each race. During a race, however, teams are unable to alter the car, so they must improve their cars' finishing positions via race strategy, i.e. optimising their selection of which tyre compounds to put on the car and when to do so. In this work, we introduce a reinforcement learning model, RSRL (Race Strategy Reinforcement Learning), to control race strategies in simulations, offering a faster alternative to the industry standard of hard-coded and Monte Carlo-based race strategies. Controlling cars with a pace equating to an expected finishing position of P5.5 (where P1 represents first place and P20 is last place), RSRL achieves an average finishing position of P5.33 on our test race, the 2023 Bahrain Grand Prix, outperforming the best baseline of P5.63. We then demonstrate, in a generalisability study, how performance for one track or multiple tracks can be prioritised via training. Further, we supplement model predictions with feature importance, decision tree-based surrogate models, and decision tree counterfactuals towards improving user trust in the model. Finally, we provide illustrations which exemplify our approach in real-world situations, drawing parallels between simulations and reality.
2501.04070
More is not always better? Enhancing Many-Shot In-Context Learning with Differentiated and Reweighting Objectives
cs.LG cs.AI cs.CL
Large language models (LLMs) excel at few-shot in-context learning (ICL) without requiring parameter updates. However, as the number of ICL demonstrations increases from a few to many, performance tends to plateau and eventually decline. We identify two primary causes for this trend: the suboptimal negative log-likelihood (NLL) optimization objective and the incremental data noise. To address these issues, we introduce DrICL, a novel optimization method that enhances model performance through Differentiated Learning and advantage-based Reweighting objectives. Globally, DrICL utilizes differentiated learning to optimize the NLL objective, ensuring that many-shot performance surpasses zero-shot levels. Locally, it dynamically adjusts the weighting of many-shot demonstrations by leveraging cumulative advantages inspired by reinforcement learning, thereby improving generalization. This approach allows the model to handle varying numbers of shots effectively, mitigating the impact of noisy data. Recognizing the lack of multi-task datasets with diverse many-shot distributions, we develop the Many-Shot ICL Benchmark (ICL-50)-a large-scale benchmark of 50 tasks that cover shot numbers from 1 to 350 within sequences of up to 8,000 tokens-for fine-tuning purposes. ICL-50 facilitates the evaluation of many-shot ICL strategies across seven prominent NLP tasks and 50 distinct datasets. Experimental results demonstrate that LLMs enhanced with DrICL achieve significant improvements in many-shot setups across various tasks, including both in-domain and out-of-domain scenarios. We release the code and benchmark dataset hoping to facilitate further research in many-shot ICL.
2501.04072
Multi-armed Bandit and Backbone boost Lin-Kernighan-Helsgaun Algorithm for the Traveling Salesman Problems
cs.DS cs.AI
The Lin-Kernighan-Helsguan (LKH) heuristic is a classic local search algorithm for the Traveling Salesman Problem (TSP). LKH introduces an $\alpha$-value to replace the traditional distance metric for evaluating the edge quality, which leads to a significant improvement. However, we observe that the $\alpha$-value does not make full use of the historical information during the search, and single guiding information often makes LKH hard to escape from some local optima. To address the above issues, we propose a novel way to extract backbone information during the TSP local search process, which is dynamic and can be updated once a local optimal solution is found. We further propose to combine backbone information, $\alpha$-value, and distance to evaluate the edge quality so as to guide the search. Moreover, we abstract their different combinations to arms in a multi-armed bandit (MAB) and use an MAB model to help the algorithm select an appropriate evaluation metric dynamically. Both the backbone information and MAB can provide diverse guiding information and learn from the search history to suggest the best metric. We apply our methods to LKH and LKH-3, which is an extension version of LKH that can be used to solve about 40 variant problems of TSP and Vehicle Routing Problem (VRP). Extensive experiments show the excellent performance and generalization capability of our proposed method, significantly improving LKH for TSP and LKH-3 for two representative TSP and VRP variants, the Colored TSP (CTSP) and Capacitated VRP with Time Windows (CVRPTW).
2501.04073
Deep Learning for Ophthalmology: The State-of-the-Art and Future Trends
eess.IV cs.CV
The emergence of artificial intelligence (AI), particularly deep learning (DL), has marked a new era in the realm of ophthalmology, offering transformative potential for the diagnosis and treatment of posterior segment eye diseases. This review explores the cutting-edge applications of DL across a range of ocular conditions, including diabetic retinopathy, glaucoma, age-related macular degeneration, and retinal vessel segmentation. We provide a comprehensive overview of foundational ML techniques and advanced DL architectures, such as CNNs, attention mechanisms, and transformer-based models, highlighting the evolving role of AI in enhancing diagnostic accuracy, optimizing treatment strategies, and improving overall patient care. Additionally, we present key challenges in integrating AI solutions into clinical practice, including ensuring data diversity, improving algorithm transparency, and effectively leveraging multimodal data. This review emphasizes AI's potential to improve disease diagnosis and enhance patient care while stressing the importance of collaborative efforts to overcome these barriers and fully harness AI's impact in advancing eye care.
2501.04074
NeRFs are Mirror Detectors: Using Structural Similarity for Multi-View Mirror Scene Reconstruction with 3D Surface Primitives
cs.CV
While neural radiance fields (NeRF) led to a breakthrough in photorealistic novel view synthesis, handling mirroring surfaces still denotes a particular challenge as they introduce severe inconsistencies in the scene representation. Previous attempts either focus on reconstructing single reflective objects or rely on strong supervision guidance in terms of additional user-provided annotations of visible image regions of the mirrors, thereby limiting the practical usability. In contrast, in this paper, we present NeRF-MD, a method which shows that NeRFs can be considered as mirror detectors and which is capable of reconstructing neural radiance fields of scenes containing mirroring surfaces without the need for prior annotations. To this end, we first compute an initial estimate of the scene geometry by training a standard NeRF using a depth reprojection loss. Our key insight lies in the fact that parts of the scene corresponding to a mirroring surface will still exhibit a significant photometric inconsistency, whereas the remaining parts are already reconstructed in a plausible manner. This allows us to detect mirror surfaces by fitting geometric primitives to such inconsistent regions in this initial stage of the training. Using this information, we then jointly optimize the radiance field and mirror geometry in a second training stage to refine their quality. We demonstrate the capability of our method to allow the faithful detection of mirrors in the scene as well as the reconstruction of a single consistent scene representation, and demonstrate its potential in comparison to baseline and mirror-aware approaches.
2501.04099
Neighbor displacement-based enhanced synthetic oversampling for multiclass imbalanced data
cs.LG
Imbalanced multiclass datasets pose challenges for machine learning algorithms. These datasets often contain minority classes that are important for accurate prediction. Existing methods still suffer from sparse data and may not accurately represent the original data patterns, leading to noise and poor model performance. A hybrid method called Neighbor Displacement-based Enhanced Synthetic Oversampling (NDESO) is proposed in this paper. This approach uses a displacement strategy for noisy data points, computing the average distance to their neighbors and moving them closer to their centroids. Random oversampling is then performed to achieve dataset balance. Extensive evaluations compare 14 alternatives on nine classifiers across synthetic and 20 real-world datasets with varying imbalance ratios. The results show that our method outperforms its competitors regarding average G-mean score and achieves the lowest statistical mean rank. This highlights its superiority and suitability for addressing data imbalance in practical applications.
2501.04102
Enhancing Distribution and Label Consistency for Graph Out-of-Distribution Generalization
cs.LG cs.AI
To deal with distribution shifts in graph data, various graph out-of-distribution (OOD) generalization techniques have been recently proposed. These methods often employ a two-step strategy that first creates augmented environments and subsequently identifies invariant subgraphs to improve generalizability. Nevertheless, this approach could be suboptimal from the perspective of consistency. First, the process of augmenting environments by altering the graphs while preserving labels may lead to graphs that are not realistic or meaningfully related to the origin distribution, thus lacking distribution consistency. Second, the extracted subgraphs are obtained from directly modifying graphs, and may not necessarily maintain a consistent predictive relationship with their labels, thereby impacting label consistency. In response to these challenges, we introduce an innovative approach that aims to enhance these two types of consistency for graph OOD generalization. We propose a modifier to obtain both augmented and invariant graphs in a unified manner. With the augmented graphs, we enrich the training data without compromising the integrity of label-graph relationships. The label consistency enhancement in our framework further preserves the supervision information in the invariant graph. We conduct extensive experiments on real-world datasets to demonstrate the superiority of our framework over other state-of-the-art baselines.
2501.04104
Security by Design Issues in Autonomous Vehicles
eess.SY cs.CR cs.SY
As autonomous vehicle (AV) technology advances towards maturity, it becomes imperative to examine the security vulnerabilities within these cyber-physical systems. While conventional cyber-security concerns are often at the forefront of discussions, it is essential to get deeper into the various layers of vulnerability that are often overlooked within mainstream frameworks. Our goal is to spotlight imminent challenges faced by AV operators and explore emerging technologies for comprehensive solutions. This research outlines the diverse security layers, spanning physical, cyber, coding, and communication aspects, in the context of AVs. Furthermore, we provide insights into potential solutions for each potential attack vector, ensuring that autonomous vehicles remain secure and resilient in an evolving threat landscape.
2501.04105
DeepVIVONet: Using deep neural operators to optimize sensor locations with application to vortex-induced vibrations
cs.LG math.OC physics.flu-dyn
We introduce DeepVIVONet, a new framework for optimal dynamic reconstruction and forecasting of the vortex-induced vibrations (VIV) of a marine riser, using field data. We demonstrate the effectiveness of DeepVIVONet in accurately reconstructing the motion of an off--shore marine riser by using sparse spatio-temporal measurements. We also show the generalization of our model in extrapolating to other flow conditions via transfer learning, underscoring its potential to streamline operational efficiency and enhance predictive accuracy. The trained DeepVIVONet serves as a fast and accurate surrogate model for the marine riser, which we use in an outer--loop optimization algorithm to obtain the optimal locations for placing the sensors. Furthermore, we employ an existing sensor placement method based on proper orthogonal decomposition (POD) to compare with our data-driven approach. We find that that while POD offers a good approach for initial sensor placement, DeepVIVONet's adaptive capabilities yield more precise and cost-effective configurations.
2501.04108
TrojanDec: Data-free Detection of Trojan Inputs in Self-supervised Learning
cs.CR cs.AI
An image encoder pre-trained by self-supervised learning can be used as a general-purpose feature extractor to build downstream classifiers for various downstream tasks. However, many studies showed that an attacker can embed a trojan into an encoder such that multiple downstream classifiers built based on the trojaned encoder simultaneously inherit the trojan behavior. In this work, we propose TrojanDec, the first data-free method to identify and recover a test input embedded with a trigger. Given a (trojaned or clean) encoder and a test input, TrojanDec first predicts whether the test input is trojaned. If not, the test input is processed in a normal way to maintain the utility. Otherwise, the test input will be further restored to remove the trigger. Our extensive evaluation shows that TrojanDec can effectively identify the trojan (if any) from a given test input and recover it under state-of-the-art trojan attacks. We further demonstrate by experiments that our TrojanDec outperforms the state-of-the-art defenses.