id
stringlengths
9
16
title
stringlengths
4
278
categories
stringlengths
5
104
abstract
stringlengths
6
4.09k
2501.01761
Adverse Weather Conditions Augmentation of LiDAR Scenes with Latent Diffusion Models
cs.CV
LiDAR scenes constitute a fundamental source for several autonomous driving applications. Despite the existence of several datasets, scenes from adverse weather conditions are rarely available. This limits the robustness of downstream machine learning models, and restrains the reliability of autonomous driving systems in particular locations and seasons. Collecting feature-diverse scenes under adverse weather conditions is challenging due to seasonal limitations. Generative models are therefore essentials, especially for generating adverse weather conditions for specific driving scenarios. In our work, we propose a latent diffusion process constituted by autoencoder and latent diffusion models. Moreover, we leverage the clear condition LiDAR scenes with a postprocessing step to improve the realism of the generated adverse weather condition scenes.
2501.01763
Quantifying A Firm's AI Engagement: Constructing Objective, Data-Driven, AI Stock Indices Using 10-K Filings
q-fin.GN cs.AI econ.EM q-fin.PM q-fin.RM
Following an analysis of existing AI-related exchange-traded funds (ETFs), we reveal the selection criteria for determining which stocks qualify as AI-related are often opaque and rely on vague phrases and subjective judgments. This paper proposes a new, objective, data-driven approach using natural language processing (NLP) techniques to classify AI stocks by analyzing annual 10-K filings from 3,395 NASDAQ-listed firms between 2011 and 2023. This analysis quantifies each company's engagement with AI through binary indicators and weighted AI scores based on the frequency and context of AI-related terms. Using these metrics, we construct four AI stock indices-the Equally Weighted AI Index (AII), the Size-Weighted AI Index (SAII), and two Time-Discounted AI Indices (TAII05 and TAII5X)-offering different perspectives on AI investment. We validate our methodology through an event study on the launch of OpenAI's ChatGPT, demonstrating that companies with higher AI engagement saw significantly greater positive abnormal returns, with analyses supporting the predictive power of our AI measures. Our indices perform on par with or surpass 14 existing AI-themed ETFs and the Nasdaq Composite Index in risk-return profiles, market responsiveness, and overall performance, achieving higher average daily returns and risk-adjusted metrics without increased volatility. These results suggest our NLP-based approach offers a reliable, market-responsive, and cost-effective alternative to existing AI-related ETF products. Our innovative methodology can also guide investors, asset managers, and policymakers in using corporate data to construct other thematic portfolios, contributing to a more transparent, data-driven, and competitive approach.
2501.01765
SaLoRA: Safety-Alignment Preserved Low-Rank Adaptation
cs.LG
As advancements in large language models (LLMs) continue and the demand for personalized models increases, parameter-efficient fine-tuning (PEFT) methods (e.g., LoRA) will become essential due to their efficiency in reducing computation costs. However, recent studies have raised alarming concerns that LoRA fine-tuning could potentially compromise the safety alignment in LLMs, posing significant risks for the model owner. In this paper, we first investigate the underlying mechanism by analyzing the changes in safety alignment related features before and after fine-tuning. Then, we propose a fixed safety module calculated by safety data and a task-specific initialization for trainable parameters in low-rank adaptations, termed Safety-alignment preserved Low-Rank Adaptation (SaLoRA). Unlike previous LoRA methods and their variants, SaLoRA enables targeted modifications to LLMs without disrupting their original alignments. Our experiments show that SaLoRA outperforms various adapters-based approaches across various evaluation metrics in different fine-tuning tasks.
2501.01767
LogicAD: Explainable Anomaly Detection via VLM-based Text Feature Extraction
cs.CV
Logical image understanding involves interpreting and reasoning about the relationships and consistency within an image's visual content. This capability is essential in applications such as industrial inspection, where logical anomaly detection is critical for maintaining high-quality standards and minimizing costly recalls. Previous research in anomaly detection (AD) has relied on prior knowledge for designing algorithms, which often requires extensive manual annotations, significant computing power, and large amounts of data for training. Autoregressive, multimodal Vision Language Models (AVLMs) offer a promising alternative due to their exceptional performance in visual reasoning across various domains. Despite this, their application to logical AD remains unexplored. In this work, we investigate using AVLMs for logical AD and demonstrate that they are well-suited to the task. Combining AVLMs with format embedding and a logic reasoner, we achieve SOTA performance on public benchmarks, MVTec LOCO AD, with an AUROC of 86.0% and F1-max of 83.7%, along with explanations of anomalies. This significantly outperforms the existing SOTA method by a large margin.
2501.01770
TCPFormer: Learning Temporal Correlation with Implicit Pose Proxy for 3D Human Pose Estimation
cs.CV
Recent multi-frame lifting methods have dominated the 3D human pose estimation. However, previous methods ignore the intricate dependence within the 2D pose sequence and learn single temporal correlation. To alleviate this limitation, we propose TCPFormer, which leverages an implicit pose proxy as an intermediate representation. Each proxy within the implicit pose proxy can build one temporal correlation therefore helping us learn more comprehensive temporal correlation of human motion. Specifically, our method consists of three key components: Proxy Update Module (PUM), Proxy Invocation Module (PIM), and Proxy Attention Module (PAM). PUM first uses pose features to update the implicit pose proxy, enabling it to store representative information from the pose sequence. PIM then invocates and integrates the pose proxy with the pose sequence to enhance the motion semantics of each pose. Finally, PAM leverages the above mapping between the pose sequence and pose proxy to enhance the temporal correlation of the whole pose sequence. Experiments on the Human3.6M and MPI-INF-3DHP datasets demonstrate that our proposed TCPFormer outperforms the previous state-of-the-art methods.
2501.01773
Compressed Domain Prior-Guided Video Super-Resolution for Cloud Gaming Content
eess.IV cs.CV
Cloud gaming is an advanced form of Internet service that necessitates local terminals to decode within limited resources and time latency. Super-Resolution (SR) techniques are often employed on these terminals as an efficient way to reduce the required bit-rate bandwidth for cloud gaming. However, insufficient attention has been paid to SR of compressed game video content. Most SR networks amplify block artifacts and ringing effects in decoded frames while ignoring edge details of game content, leading to unsatisfactory reconstruction results. In this paper, we propose a novel lightweight network called Coding Prior-Guided Super-Resolution (CPGSR) to address the SR challenges in compressed game video content. First, we design a Compressed Domain Guided Block (CDGB) to extract features of different depths from coding priors, which are subsequently integrated with features from the U-net backbone. Then, a series of re-parameterization blocks are utilized for reconstruction. Ultimately, inspired by the quantization in video coding, we propose a partitioned focal frequency loss to effectively guide the model's focus on preserving high-frequency information. Extensive experiments demonstrate the advancement of our approach.
2501.01774
A Unifying View of Linear Function Approximation in Off-Policy RL Through Matrix Splitting and Preconditioning
cs.LG
Traditionally, TD and FQI are viewed as differing in the number of updates toward the target value function: TD makes one update, FQI makes an infinite number, and Partial Fitted Q-Iteration (PFQI) performs a finite number, such as the use of a target network in Deep Q-Networks (DQN) in the OPE setting. This perspective, however, fails to capture the convergence connections between these algorithms and may lead to incorrect conclusions, for example, that the convergence of TD implies the convergence of FQI. In this paper, we focus on linear value function approximation and offer a new perspective, unifying TD, FQI, and PFQI as the same iterative method for solving the Least Squares Temporal Difference (LSTD) system, but using different preconditioners and matrix splitting schemes. TD uses a constant preconditioner, FQI employs a data-feature adaptive preconditioner, and PFQI transitions between the two. Then, we reveal that in the context of linear function approximation, increasing the number of updates under the same target value function essentially represents a transition from using a constant preconditioner to data-feature adaptive preconditioner. This unifying perspective also simplifies the analyses of the convergence conditions for these algorithms and clarifies many issues. Consequently, we fully characterize the convergence of each algorithm without assuming specific properties of the chosen features (e.g., linear independence). We also examine how common assumptions about feature representations affect convergence, and discover new conditions on features that are important for convergence. These convergence conditions allow us to establish the convergence connections between these algorithms and to address important questions.
2501.01776
Smooth Rate Limiter Model for Power System Stability Analysis and Control
eess.SY cs.SY
The letter proposes a smooth Rate Limiter (RL) model for power system stability analysis and control. The proposed model enables the effects of derivative bounds to be incorporated into system eigenvalue analysis, while replicating the behavior of conventional non-smooth RLs with high fidelity. In addition, it can be duly modified to enhance the system's dynamic control performance. The behavior of the proposed model is demonstrated through illustrative examples as well as through a simulation of the New York/New England 16-machine 68-bus system.
2501.01779
From Occasional to Steady: Habit Formation Insights From a Comprehensive Fitness Study
cs.CY cs.CE cs.SI
Exercising regularly is widely recognized as a cornerstone of health, yet the challenge of sustaining consistent exercise habits persists. Understanding the factors that influence the formation of these habits is crucial for developing effective interventions. This study utilizes data from Mars Athletic Club, T\"urkiye's largest sports chain, to investigate the dynamics of gym attendance and habit formation. The general problem addressed by this study is identifying the critical periods and factors that contribute to the successful establishment of consistent exercise routines among gym-goers. Here we show that there are specific periods during which gym attendance is most crucial for habit formation. By developing a survival metric based on gym attendance patterns, we pinpoint these critical periods and segment members into distinct clusters based on their visit patterns. Our analysis reveals significant differences in how various subgroups respond to interventions, such as group classes, personal trainer sessions, and visiting different clubs. Using causal inference analysis, we demonstrate that personalized guidance and social dynamics are key drivers of sustained long-term engagement. By systematically examining these variables and considering the specific characteristics of different clusters, our research demonstrates the importance of a tailored, multi-dimensional approach to promoting exercise habits, which integrates social dynamics, personalized guidance, and strategic interventions to sustain long-term engagement.
2501.01785
Can Synthetic Data be Fair and Private? A Comparative Study of Synthetic Data Generation and Fairness Algorithms
cs.LG cs.AI cs.CY
The increasing use of machine learning in learning analytics (LA) has raised significant concerns around algorithmic fairness and privacy. Synthetic data has emerged as a dual-purpose tool, enhancing privacy and improving fairness in LA models. However, prior research suggests an inverse relationship between fairness and privacy, making it challenging to optimize both. This study investigates which synthetic data generators can best balance privacy and fairness, and whether pre-processing fairness algorithms, typically applied to real datasets, are effective on synthetic data. Our results highlight that the DEbiasing CAusal Fairness (DECAF) algorithm achieves the best balance between privacy and fairness. However, DECAF suffers in utility, as reflected in its predictive accuracy. Notably, we found that applying pre-processing fairness algorithms to synthetic data improves fairness even more than when applied to real data. These findings suggest that combining synthetic data generation with fairness pre-processing offers a promising approach to creating fairer LA models.
2501.01788
Universal Online Temporal Calibration for Optimization-based Visual-Inertial Navigation Systems
cs.RO cs.CV
6-Degree of Freedom (6DoF) motion estimation with a combination of visual and inertial sensors is a growing area with numerous real-world applications. However, precise calibration of the time offset between these two sensor types is a prerequisite for accurate and robust tracking. To address this, we propose a universal online temporal calibration strategy for optimization-based visual-inertial navigation systems. Technically, we incorporate the time offset td as a state parameter in the optimization residual model to align the IMU state to the corresponding image timestamp using td, angular velocity and translational velocity. This allows the temporal misalignment td to be optimized alongside other tracking states during the process. As our method only modifies the structure of the residual model, it can be applied to various optimization-based frameworks with different tracking frontends. We evaluate our calibration method with both EuRoC and simulation data and extensive experiments demonstrate that our approach provides more accurate time offset estimation and faster convergence, particularly in the presence of noisy sensor data.
2501.01790
Ingredients: Blending Custom Photos with Video Diffusion Transformers
cs.CV
This paper presents a powerful framework to customize video creations by incorporating multiple specific identity (ID) photos, with video diffusion Transformers, referred to as \texttt{Ingredients}. Generally, our method consists of three primary modules: (\textbf{i}) a facial extractor that captures versatile and precise facial features for each human ID from both global and local perspectives; (\textbf{ii}) a multi-scale projector that maps face embeddings into the contextual space of image query in video diffusion transformers; (\textbf{iii}) an ID router that dynamically combines and allocates multiple ID embedding to the corresponding space-time regions. Leveraging a meticulously curated text-video dataset and a multi-stage training protocol, \texttt{Ingredients} demonstrates superior performance in turning custom photos into dynamic and personalized video content. Qualitative evaluations highlight the advantages of proposed method, positioning it as a significant advancement toward more effective generative video control tools in Transformer-based architecture, compared to existing methods. The data, code, and model weights are publicly available at: \url{https://github.com/feizc/Ingredients}.
2501.01791
A Minimal Subset Approach for Efficient and Scalable Loop Closure
cs.CV cs.RO
Loop closure detection in large-scale and long-term missions can be computationally demanding due to the need to identify, verify, and process numerous candidate pairs to establish edge connections for the pose graph optimization. Keyframe sampling mitigates this by reducing the number of frames stored and processed in the back-end system. In this article, we address the gap in optimized keyframe sampling for the combined problem of pose graph optimization and loop closure detection. Our Minimal Subset Approach (MSA) employs an optimization strategy with two key factors, redundancy minimization and information preservation, within a sliding window framework to efficiently reduce redundant keyframes, while preserving essential information. This method delivers comparable performance to baseline approaches, while enhancing scalability and reducing computational overhead. Finally, we evaluate MSA on relevant publicly available datasets, showcasing that it consistently performs across a wide range of environments, without requiring any manual parameter tuning.
2501.01793
Creating Artificial Students that Never Existed: Leveraging Large Language Models and CTGANs for Synthetic Data Generation
cs.LG cs.AI
In this study, we explore the growing potential of AI and deep learning technologies, particularly Generative Adversarial Networks (GANs) and Large Language Models (LLMs), for generating synthetic tabular data. Access to quality students data is critical for advancing learning analytics, but privacy concerns and stricter data protection regulations worldwide limit their availability and usage. Synthetic data offers a promising alternative. We investigate whether synthetic data can be leveraged to create artificial students for serving learning analytics models. Using the popular GAN model CTGAN and three LLMs- GPT2, DistilGPT2, and DialoGPT, we generate synthetic tabular student data. Our results demonstrate the strong potential of these methods to produce high-quality synthetic datasets that resemble real students data. To validate our findings, we apply a comprehensive set of utility evaluation metrics to assess the statistical and predictive performance of the synthetic data and compare the different generator models used, specially the performance of LLMs. Our study aims to provide the learning analytics community with valuable insights into the use of synthetic data, laying the groundwork for expanding the field methodological toolbox with new innovative approaches for learning analytics data generation.
2501.01796
Reading Between the Lines: A dataset and a study on why some texts are tougher than others
cs.CL
Our research aims at better understanding what makes a text difficult to read for specific audiences with intellectual disabilities, more specifically, people who have limitations in cognitive functioning, such as reading and understanding skills, an IQ below 70, and challenges in conceptual domains. We introduce a scheme for the annotation of difficulties which is based on empirical research in psychology as well as on research in translation studies. The paper describes the annotated dataset, primarily derived from the parallel texts (standard English and Easy to Read English translations) made available online. we fine-tuned four different pre-trained transformer models to perform the task of multiclass classification to predict the strategies required for simplification. We also investigate the possibility to interpret the decisions of this language model when it is aimed at predicting the difficulty of sentences. The resources are available from https://github.com/Nouran-Khallaf/why-tough
2501.01798
JoyGen: Audio-Driven 3D Depth-Aware Talking-Face Video Editing
cs.CV
Significant progress has been made in talking-face video generation research; however, precise lip-audio synchronization and high visual quality remain challenging in editing lip shapes based on input audio. This paper introduces JoyGen, a novel two-stage framework for talking-face generation, comprising audio-driven lip motion generation and visual appearance synthesis. In the first stage, a 3D reconstruction model and an audio2motion model predict identity and expression coefficients respectively. Next, by integrating audio features with a facial depth map, we provide comprehensive supervision for precise lip-audio synchronization in facial generation. Additionally, we constructed a Chinese talking-face dataset containing 130 hours of high-quality video. JoyGen is trained on the open-source HDTF dataset and our curated dataset. Experimental results demonstrate superior lip-audio synchronization and visual quality achieved by our method.
2501.01799
Grasping in Uncertain Environments: A Case Study For Industrial Robotic Recycling
cs.RO cs.SY eess.SY
Autonomous robotic grasping of uncertain objects in uncertain environments is an impactful open challenge for the industries of the future. One such industry is the recycling of Waste Electrical and Electronic Equipment (WEEE) materials, in which electric devices are disassembled and readied for the recovery of raw materials. Since devices may contain hazardous materials and their disassembly involves heavy manual labor, robotic disassembly is a promising venue. However, since devices may be damaged, dirty and unidentified, robotic disassembly is challenging since object models are unavailable or cannot be relied upon. This case study explores grasping strategies for industrial robotic disassembly of WEEE devices with uncertain vision data. We propose three grippers and appropriate tactile strategies for force-based manipulation that improves grasping robustness. For each proposed gripper, we develop corresponding strategies that can perform effectively in different grasping tasks and leverage the grippers design and unique strengths. Through experiments conducted in lab and factory settings for four different WEEE devices, we demonstrate how object uncertainty may be overcome by tactile sensing and compliant techniques, significantly increasing grasping success rates.
2501.01801
John Ellipsoids via Lazy Updates
cs.DS cs.LG
We give a faster algorithm for computing an approximate John ellipsoid around $n$ points in $d$ dimensions. The best known prior algorithms are based on repeatedly computing the leverage scores of the points and reweighting them by these scores [CCLY19]. We show that this algorithm can be substantially sped up by delaying the computation of high accuracy leverage scores by using sampling, and then later computing multiple batches of high accuracy leverage scores via fast rectangular matrix multiplication. We also give low-space streaming algorithms for John ellipsoids using similar ideas.
2501.01802
BERT4MIMO: A Foundation Model using BERT Architecture for Massive MIMO Channel State Information Prediction
cs.IT cs.AI eess.SP math.IT
Massive MIMO (Multiple-Input Multiple-Output) is an advanced wireless communication technology, using a large number of antennas to improve the overall performance of the communication system in terms of capacity, spectral, and energy efficiency. The performance of MIMO systems is highly dependent on the quality of channel state information (CSI). Predicting CSI is, therefore, essential for improving communication system performance, particularly in MIMO systems, since it represents key characteristics of a wireless channel, including propagation, fading, scattering, and path loss. This study proposes a foundation model inspired by BERT, called BERT4MIMO, which is specifically designed to process high-dimensional CSI data from massive MIMO systems. BERT4MIMO offers superior performance in reconstructing CSI under varying mobility scenarios and channel conditions through deep learning and attention mechanisms. The experimental results demonstrate the effectiveness of BERT4MIMO in a variety of wireless environments.
2501.01805
End-to-End Long Document Summarization using Gradient Caching
cs.CL cs.AI
Training transformer-based encoder-decoder models for long document summarization poses a significant challenge due to the quadratic memory consumption during training. Several approaches have been proposed to extend the input length at test time, but training with these approaches is still difficult, requiring truncation of input documents and causing a mismatch between training and test conditions. In this work, we propose CachED (Gradient $\textbf{Cach}$ing for $\textbf{E}$ncoder-$\textbf{D}$ecoder models), an approach that enables end-to-end training of existing transformer-based encoder-decoder models, using the entire document without truncation. Specifically, we apply non-overlapping sliding windows to input documents, followed by fusion in decoder. During backpropagation, the gradients are cached at the decoder and are passed through the encoder in chunks by re-computing the hidden vectors, similar to gradient checkpointing. In the experiments on long document summarization, we extend BART to CachED BART, processing more than 500K tokens during training and achieving superior performance without using any additional parameters.
2501.01806
TRG-planner: Traversal Risk Graph-Based Path Planning in Unstructured Environments for Safe and Efficient Navigation
cs.RO cs.SY eess.SY
Unstructured environments such as mountains, caves, construction sites, or disaster areas are challenging for autonomous navigation because of terrain irregularities. In particular, it is crucial to plan a path to avoid risky terrain and reach the goal quickly and safely. In this paper, we propose a method for safe and distance-efficient path planning, leveraging Traversal Risk Graph (TRG), a novel graph representation that takes into account geometric traversability of the terrain. TRG nodes represent stability and reachability of the terrain, while edges represent relative traversal risk-weighted path candidates. Additionally, TRG is constructed in a wavefront propagation manner and managed hierarchically, enabling real-time planning even in large-scale environments. Lastly, we formulate a graph optimization problem on TRG that leads the robot to navigate by prioritizing both safe and short paths. Our approach demonstrated superior safety, distance efficiency, and fast processing time compared to the conventional methods. It was also validated in several real-world experiments using a quadrupedal robot. Notably, TRG-planner contributed as the global path planner of an autonomous navigation framework for the DreamSTEP team, which won the Quadruped Robot Challenge at ICRA 2023. The project page is available at https://trg-planner.github.io .
2501.01808
MoEE: Mixture of Emotion Experts for Audio-Driven Portrait Animation
cs.CV
The generation of talking avatars has achieved significant advancements in precise audio synchronization. However, crafting lifelike talking head videos requires capturing a broad spectrum of emotions and subtle facial expressions. Current methods face fundamental challenges: a) the absence of frameworks for modeling single basic emotional expressions, which restricts the generation of complex emotions such as compound emotions; b) the lack of comprehensive datasets rich in human emotional expressions, which limits the potential of models. To address these challenges, we propose the following innovations: 1) the Mixture of Emotion Experts (MoEE) model, which decouples six fundamental emotions to enable the precise synthesis of both singular and compound emotional states; 2) the DH-FaceEmoVid-150 dataset, specifically curated to include six prevalent human emotional expressions as well as four types of compound emotions, thereby expanding the training potential of emotion-driven models. Furthermore, to enhance the flexibility of emotion control, we propose an emotion-to-latents module that leverages multimodal inputs, aligning diverse control signals-such as audio, text, and labels-to ensure more varied control inputs as well as the ability to control emotions using audio alone. Through extensive quantitative and qualitative evaluations, we demonstrate that the MoEE framework, in conjunction with the DH-FaceEmoVid-150 dataset, excels in generating complex emotional expressions and nuanced facial details, setting a new benchmark in the field. These datasets will be publicly released.
2501.01811
QuantumBind-RBFE: Accurate Relative Binding Free Energy Calculations Using Neural Network Potentials
physics.chem-ph cs.LG physics.comp-ph
Accurate prediction of protein-ligand binding affinities is crucial in drug discovery, particularly during hit-to-lead and lead optimization phases, however, limitations in ligand force fields continue to impact prediction accuracy. In this work, we validate relative binding free energy (RBFE) accuracy using neural network potentials (NNPs) for the ligands. We utilize a novel NNP model, AceForce 1.0, based on the TensorNet architecture for small molecules that broadens the applicability to diverse drug-like compounds, including all important chemical elements and supporting charged molecules. Using established benchmarks, we show overall improved accuracy and correlation in binding affinity predictions compared with GAFF2 for molecular mechanics and ANI2-x for NNPs. Slightly less accuracy but comparable correlations with OPLS4. We also show that we can run the NNP simulations at 2 fs timestep, at least two times larger than previous NNP models, providing significant speed gains. The results show promise for further evolutions of free energy calculations using NNPs while demonstrating its practical use already with the current generation. The code and NNP model are publicly available for research use.
2501.01813
Eliciting Understandable Architectonic Gestures for Robotic Furniture through Co-Design Improvisation
cs.HC cs.RO
The vision of adaptive architecture proposes that robotic technologies could enable interior spaces to physically transform in a bidirectional interaction with occupants. Yet, it is still unknown how this interaction could unfold in an understandable way. Inspired by HRI studies where robotic furniture gestured intents to occupants by deliberately positioning or moving in space, we hypothesise that adaptive architecture could also convey intents through gestures performed by a mobile robotic partition. To explore this design space, we invited 15 multidisciplinary experts to join co-design improvisation sessions, where they manually manoeuvred a deactivated robotic partition to design gestures conveying six architectural intents that varied in purpose and urgency. Using a gesture elicitation method alongside motion-tracking data, a Laban-based questionnaire, and thematic analysis, we identified 20 unique gestural strategies. Through categorisation, we introduced architectonic gestures as a novel strategy for robotic furniture to convey intent by indexically leveraging its spatial impact, complementing the established deictic and emblematic gestures. Our study thus represents an exploratory step toward making the autonomous gestures of adaptive architecture more legible. By understanding how robotic gestures are interpreted based not only on their motion but also on their spatial impact, we contribute to bridging HRI with Human-Building Interaction research.
2501.01816
Uncertainty-Aware Label Refinement on Hypergraphs for Personalized Federated Facial Expression Recognition
cs.CV
Most facial expression recognition (FER) models are trained on large-scale expression data with centralized learning. Unfortunately, collecting a large amount of centralized expression data is difficult in practice due to privacy concerns of facial images. In this paper, we investigate FER under the framework of personalized federated learning, which is a valuable and practical decentralized setting for real-world applications. To this end, we develop a novel uncertainty-Aware label refineMent on hYpergraphs (AMY) method. For local training, each local model consists of a backbone, an uncertainty estimation (UE) block, and an expression classification (EC) block. In the UE block, we leverage a hypergraph to model complex high-order relationships between expression samples and incorporate these relationships into uncertainty features. A personalized uncertainty estimator is then introduced to estimate reliable uncertainty weights of samples in the local client. In the EC block, we perform label propagation on the hypergraph, obtaining high-quality refined labels for retraining an expression classifier. Based on the above, we effectively alleviate heterogeneous sample uncertainty across clients and learn a robust personalized FER model in each client. Experimental results on two challenging real-world facial expression databases show that our proposed method consistently outperforms several state-of-the-art methods. This indicates the superiority of hypergraph modeling for uncertainty estimation and label refinement on the personalized federated FER task. The source code will be released at https://github.com/mobei1006/AMY.
2501.01817
Distributed Framework Construction for Affine Formation Control
eess.SY cs.SY
In affine formation control problems, the construction of the framework with universal rigidity and affine localizability is a critical prerequisite, but it has not yet been well addressed, especially when additional agents join the formation or link/agent failures emerge. Motivated by this observation, we investigate the problem of constructing affine formation frameworks in three scenarios, including vertex addition, edge deletion and vertex deletion. Our approach starts from the original affine formation and uses geometric methods to locally adjust the structure of the weighted graph to describe the topology, so that the modified framework maintains the universal rigidity and affine localizability. Notably, the developed strategies only utilize local measurements and exhibit distributed characteristics, laying the foundation for applications in multi-agent systems. To demonstrate the compatibility with affine formation control proposals, we present a case study on affine formation tracking in a multi-UAV formation, demonstrating the effectiveness of our algorithms in constructing eligible frameworks in aforementioned scenarios. Moreover, a comparative simulations is also conducted to highlight the low time complexity of our distributed algorithm relative to the centralized optimization-based method.
2501.01818
Rerouting LLM Routers
cs.CR cs.LG
LLM routers aim to balance quality and cost of generation by classifying queries and routing them to a cheaper or more expensive LLM depending on their complexity. Routers represent one type of what we call LLM control planes: systems that orchestrate use of one or more LLMs. In this paper, we investigate routers' adversarial robustness. We first define LLM control plane integrity, i.e., robustness of LLM orchestration to adversarial inputs, as a distinct problem in AI safety. Next, we demonstrate that an adversary can generate query-independent token sequences we call ``confounder gadgets'' that, when added to any query, cause LLM routers to send the query to a strong LLM. Our quantitative evaluation shows that this attack is successful both in white-box and black-box settings against a variety of open-source and commercial routers, and that confounding queries do not affect the quality of LLM responses. Finally, we demonstrate that gadgets can be effective while maintaining low perplexity, thus perplexity-based filtering is not an effective defense. We finish by investigating alternative defenses.
2501.01821
SDPO: Segment-Level Direct Preference Optimization for Social Agents
cs.AI cs.CL
Social agents powered by large language models (LLMs) can simulate human social behaviors but fall short in handling complex goal-oriented social dialogues. Direct Preference Optimization (DPO) has proven effective in aligning LLM behavior with human preferences across a variety of agent tasks. Existing DPO-based approaches for multi-turn interactions are divided into turn-level and session-level methods. The turn-level method is overly fine-grained, focusing exclusively on individual turns, while session-level methods are too coarse-grained, often introducing training noise. To address these limitations, we propose Segment-Level Direct Preference Optimization (SDPO), which focuses on specific key segments within interactions to optimize multi-turn agent behavior while minimizing training noise. Evaluations on the SOTOPIA benchmark demonstrate that SDPO-tuned agents consistently outperform both existing DPO-based methods and proprietary LLMs like GPT-4o, underscoring SDPO's potential to advance the social intelligence of LLM-based agents. We release our code and data at https://github.com/AlibabaResearch/DAMO-ConvAI/tree/main/SDPO.
2501.01825
Unified Native Spaces in Kernel Methods
stat.ML cs.LG
There exists a plethora of parametric models for positive definite kernels, and their use is ubiquitous in disciplines as diverse as statistics, machine learning, numerical analysis, and approximation theory. Usually, the kernel parameters index certain features of an associated process. Amongst those features, smoothness (in the sense of Sobolev spaces, mean square differentiability, and fractal dimensions), compact or global supports, and negative dependencies (hole effects) are of interest to several theoretical and applied disciplines. This paper unifies a wealth of well-known kernels into a single parametric class that encompasses them as special cases, attained either by exact parameterization or through parametric asymptotics. We furthermore characterize the Sobolev space that is norm equivalent to the RKHS associated with the new kernel. As a by-product, we infer the Sobolev spaces that are associated with existing classes of kernels. We illustrate the main properties of the new class, show how this class can switch from compact to global supports, and provide special cases for which the kernel attains negative values over nontrivial intervals. Hence, the proposed class of kernel is the reproducing kernel of a very rich Hilbert space that contains many special cases, including the celebrated Mat\'ern and Wendland kernels, as well as their aliases with hole effects.
2501.01827
The Proof is in the Almond Cookies
cs.CL cs.AI
This paper presents a case study on how to process cooking recipes (and more generally, how-to instructions) in a way that makes it possible for a robot or artificial cooking assistant to support human chefs in the kitchen. Such AI assistants would be of great benefit to society, as they can help to sustain the autonomy of aging adults or people with a physical impairment, or they may reduce the stress in a professional kitchen. We propose a novel approach to computational recipe understanding that mimics the human sense-making process, which is narrative-based. Using an English recipe for almond crescent cookies as illustration, we show how recipes can be modelled as rich narrative structures by integrating various knowledge sources such as language processing, ontologies, and mental simulation. We show how such narrative structures can be used for (a) dealing with the challenges of recipe language, such as zero anaphora, (b) optimizing a robot's planning process, (c) measuring how well an AI system understands its current tasks, and (d) allowing recipe annotations to become language-independent.
2501.01828
Age-Based Device Selection and Transmit Power Optimization in Over-the-Air Federated Learning
cs.NI cs.LG
Recently, over-the-air federated learning (FL) has attracted significant attention for its ability to enhance communication efficiency. However, the performance of over-the-air FL is often constrained by device selection strategies and signal aggregation errors. In particular, neglecting straggler devices in FL can lead to a decline in the fairness of model updates and amplify the global model's bias toward certain devices' data, ultimately impacting the overall system performance. To address this issue, we propose a joint device selection and transmit power optimization framework that ensures the appropriate participation of straggler devices, maintains efficient training performance, and guarantees timely updates. First, we conduct a theoretical analysis to quantify the convergence upper bound of over-the-air FL under age-of-information (AoI)-based device selection. Our analysis further reveals that both the number of selected devices and the signal aggregation errors significantly influence the convergence upper bound. To minimize the expected weighted sum peak age of information, we calculate device priorities for each communication round using Lyapunov optimization and select the highest-priority devices via a greedy algorithm. Then, we formulate and solve a transmit power and normalizing factor optimization problem for selected devices to minimize the time-average mean squared error (MSE). Experimental results demonstrate that our proposed method offers two significant advantages: (1) it reduces MSE and improves model performance compared to baseline methods, and (2) it strikes a balance between fairness and training efficiency while maintaining satisfactory timeliness, ensuring stable model performance.
2501.01830
Auto-RT: Automatic Jailbreak Strategy Exploration for Red-Teaming Large Language Models
cs.CR cs.AI cs.CL
Automated red-teaming has become a crucial approach for uncovering vulnerabilities in large language models (LLMs). However, most existing methods focus on isolated safety flaws, limiting their ability to adapt to dynamic defenses and uncover complex vulnerabilities efficiently. To address this challenge, we propose Auto-RT, a reinforcement learning framework that automatically explores and optimizes complex attack strategies to effectively uncover security vulnerabilities through malicious queries. Specifically, we introduce two key mechanisms to reduce exploration complexity and improve strategy optimization: 1) Early-terminated Exploration, which accelerate exploration by focusing on high-potential attack strategies; and 2) Progressive Reward Tracking algorithm with intermediate downgrade models, which dynamically refine the search trajectory toward successful vulnerability exploitation. Extensive experiments across diverse LLMs demonstrate that, by significantly improving exploration efficiency and automatically optimizing attack strategies, Auto-RT detects a boarder range of vulnerabilities, achieving a faster detection speed and 16.63\% higher success rates compared to existing methods.
2501.01831
Online Fault Tolerance Strategy for Abrupt Reachability Constraint Changes
eess.SY cs.RO cs.SY
When a system's constraints change abruptly, the system's reachability safety does no longer sustain. Thus, the system can reach a forbidden/dangerous value. Conventional remedy practically involves online controller redesign (OCR) to re-establish the reachability's compliance with the new constraints, which, however, is usually too slow. There is a need for an online strategy capable of managing runtime changes in reachability constraints. However, to the best of the authors' knowledge, this topic has not been addressed in the existing literature. In this paper, we propose a fast fault tolerance strategy to recover the system's reachability safety in runtime. Instead of redesigning the system's controller, we propose to change the system's reference state to modify the system's reachability to comply with the new constraints. We frame the reference state search as an optimization problem and employ the Karush-Kuhn-Tucker (KKT) method as well as the Interior Point Method (IPM) based Newton's method (as a fallback for the KKT method) for fast solution derivation. The optimization also allows more future fault tolerance. Numerical simulations demonstrate that our method outperforms the conventional OCR method in terms of computational efficiency and success rate. Specifically, the results show that the proposed method finds a solution $10^{2}$ (with the IPM based Newton's method) $\sim 10^{4}$ (with the KKT method) times faster than the OCR method. Additionally, the improvement rate of the success rate of our method over the OCR method is $40.81\%$ without considering the deadline of run time. The success rate remains at $49.44\%$ for the proposed method, while it becomes $0\%$ for the OCR method when a deadline of $1.5 \; seconds$ is imposed.
2501.01832
Time Series Language Model for Descriptive Caption Generation
cs.CL cs.LG
The automatic generation of representative natural language descriptions for observable patterns in time series data enhances interpretability, simplifies analysis and increases cross-domain utility of temporal data. While pre-trained foundation models have made considerable progress in natural language processing (NLP) and computer vision (CV), their application to time series analysis has been hindered by data scarcity. Although several large language model (LLM)-based methods have been proposed for time series forecasting, time series captioning is under-explored in the context of LLMs. In this paper, we introduce TSLM, a novel time series language model designed specifically for time series captioning. TSLM operates as an encoder-decoder model, leveraging both text prompts and time series data representations to capture subtle temporal patterns across multiple phases and generate precise textual descriptions of time series inputs. TSLM addresses the data scarcity problem in time series captioning by first leveraging an in-context prompting synthetic data generation, and second denoising the generated data via a novel cross-modal dense retrieval scoring applied to time series-caption pairs. Experimental findings on various time series captioning datasets demonstrate that TSLM outperforms existing state-of-the-art approaches from multiple data modalities by a significant margin.
2501.01834
MoColl: Agent-Based Specific and General Model Collaboration for Image Captioning
cs.CV cs.AI
Image captioning is a critical task at the intersection of computer vision and natural language processing, with wide-ranging applications across various domains. For complex tasks such as diagnostic report generation, deep learning models require not only domain-specific image-caption datasets but also the incorporation of relevant general knowledge to provide contextual accuracy. Existing approaches exhibit inherent limitations: specialized models excel in capturing domain-specific details but lack generalization, while vision-language models (VLMs) built on large language models (LLMs) leverage general knowledge but struggle with domain-specific adaptation. To address these limitations, this paper proposes a novel agent-enhanced model collaboration framework, which we call MoColl, designed to effectively integrate domain-specific and general knowledge. Specifically, our approach is to decompose complex image captioning tasks into a series of interconnected question-answer subtasks. A trainable visual question answering (VQA) model is employed as a specialized tool to focus on domain-specific visual analysis, answering task-specific questions based on image content. Concurrently, an LLM-based agent with general knowledge formulates these questions and synthesizes the resulting question-answer pairs into coherent captions. Beyond its role in leveraging the VQA model, the agent further guides its training to enhance its domain-specific capabilities. Experimental results on radiology report generation validate the effectiveness of the proposed framework, demonstrating significant improvements in the quality of generated reports.
2501.01835
ASKCOS: an open source software suite for synthesis planning
cs.AI
The advancement of machine learning and the availability of large-scale reaction datasets have accelerated the development of data-driven models for computer-aided synthesis planning (CASP) in the past decade. Here, we detail the newest version of ASKCOS, an open source software suite for synthesis planning that makes available several research advances in a freely available, practical tool. Four one-step retrosynthesis models form the basis of both interactive planning and automatic planning modes. Retrosynthetic planning is complemented by other modules for feasibility assessment and pathway evaluation, including reaction condition recommendation, reaction outcome prediction, and auxiliary capabilities such as solubility prediction and quantum mechanical descriptor prediction. ASKCOS has assisted hundreds of medicinal, synthetic, and process chemists in their day-to-day tasks, complementing expert decision making. It is our belief that CASP tools like ASKCOS are an important part of modern chemistry research, and that they offer ever-increasing utility and accessibility.
2501.01836
Practical machine learning is learning on small samples
cs.LG cs.AI
Based on limited observations, machine learning discerns a dependence which is expected to hold in the future. What makes it possible? Statistical learning theory imagines indefinitely increasing training sample to justify its approach. In reality, there is no infinite time or even infinite general population for learning. Here I argue that practical machine learning is based on an implicit assumption that underlying dependence is relatively ``smooth" : likely, there are no abrupt differences in feedback between cases with close data points. From this point of view learning shall involve selection of the hypothesis ``smoothly" approximating the training set. I formalize this as Practical learning paradigm. The paradigm includes terminology and rules for description of learners. Popular learners (local smoothing, k-NN, decision trees, Naive Bayes, SVM for classification and for regression) are shown here to be implementations of this paradigm.
2501.01840
Signal Recovery Using a Spiked Mixture Model
stat.ML cs.LG
We introduce the spiked mixture model (SMM) to address the problem of estimating a set of signals from many randomly scaled and noisy observations. Subsequently, we design a novel expectation-maximization (EM) algorithm to recover all parameters of the SMM. Numerical experiments show that in low signal-to-noise ratio regimes, and for data types where the SMM is relevant, SMM surpasses the more traditional Gaussian mixture model (GMM) in terms of signal recovery performance. The broad relevance of the SMM and its corresponding EM recovery algorithm is demonstrated by applying the technique to different data types. The first case study is a biomedical research application, utilizing an imaging mass spectrometry dataset to explore the molecular content of a rat brain tissue section at micrometer scale. The second case study demonstrates SMM performance in a computer vision application, segmenting a hyperspectral imaging dataset into underlying patterns. While the measurement modalities differ substantially, in both case studies SMM is shown to recover signals that were missed by traditional methods such as k-means clustering and GMM.
2501.01841
Dedicated Inference Engine and Binary-Weight Neural Networks for Lightweight Instance Segmentation
cs.CV cs.AR
Reducing computational costs is an important issue for development of embedded systems. Binary-weight Neural Networks (BNNs), in which weights are binarized and activations are quantized, are employed to reduce computational costs of various kinds of applications. In this paper, a design methodology of hardware architecture for inference engines is proposed to handle modern BNNs with two operation modes. Multiply-Accumulate (MAC) operations can be simplified by replacing multiply operations with bitwise operations. The proposed method can effectively reduce the gate count of inference engines by removing a part of computational costs from the hardware system. The architecture of MAC operations can calculate the inference results of BNNs efficiently with only 52% of hardware costs compared with the related works. To show that the inference engine can handle practical applications, two lightweight networks which combine the backbones of SegNeXt and the decoder of SparseInst for instance segmentation are also proposed. The output results of the lightweight networks are computed using only bitwise operations and add operations. The proposed inference engine has lower hardware costs than related works. The experimental results show that the proposed inference engine can handle the proposed instance-segmentation networks and achieves higher accuracy than YOLACT on the "Person" category although the model size is 77.7$\times$ smaller compared with YOLACT.
2501.01844
Learning from Ambiguous Data with Hard Labels
cs.LG
Real-world data often contains intrinsic ambiguity that the common single-hard-label annotation paradigm ignores. Standard training using ambiguous data with these hard labels may produce overly confident models and thus leading to poor generalization. In this paper, we propose a novel framework called Quantized Label Learning (QLL) to alleviate this issue. First, we formulate QLL as learning from (very) ambiguous data with hard labels: ideally, each ambiguous instance should be associated with a ground-truth soft-label distribution describing its corresponding probabilistic weight in each class, however, this is usually not accessible; in practice, we can only observe a quantized label, i.e., a hard label sampled (quantized) from the corresponding ground-truth soft-label distribution, of each instance, which can be seen as a biased approximation of the ground-truth soft-label. Second, we propose a Class-wise Positive-Unlabeled (CPU) risk estimator that allows us to train accurate classifiers from only ambiguous data with quantized labels. Third, to simulate ambiguous datasets with quantized labels in the real world, we design a mixing-based ambiguous data generation procedure for empirical evaluation. Experiments demonstrate that our CPU method can significantly improve model generalization performance and outperform the baselines.
2501.01845
Semantic Segmentation for Sequential Historical Maps by Learning from Only One Map
cs.CV
Historical maps are valuable resources that capture detailed geographical information from the past. However, these maps are typically available in printed formats, which are not conducive to modern computer-based analyses. Digitizing these maps into a machine-readable format enables efficient computational analysis. In this paper, we propose an automated approach to digitization using deep-learning-based semantic segmentation, which assigns a semantic label to each pixel in scanned historical maps. A key challenge in this process is the lack of ground-truth annotations required for training deep neural networks, as manual labeling is time-consuming and labor-intensive. To address this issue, we introduce a weakly-supervised age-tracing strategy for model fine-tuning. This approach exploits the similarity in appearance and land-use patterns between historical maps from neighboring time periods to guide the training process. Specifically, model predictions for one map are utilized as pseudo-labels for training on maps from adjacent time periods. Experiments conducted on our newly curated \textit{Hameln} dataset demonstrate that the proposed age-tracing strategy significantly enhances segmentation performance compared to baseline models. In the best-case scenario, the mean Intersection over Union (mIoU) achieved 77.3\%, reflecting an improvement of approximately 20\% over baseline methods. Additionally, the fine-tuned model achieved an average overall accuracy of 97\%, highlighting the effectiveness of our approach for digitizing historical maps.
2501.01849
Multi-Agent Conversational Online Learning for Adaptive LLM Response Identification
cs.HC cs.AI
The remarkable generative capability of large language models (LLMs) has sparked a growing interest in automatically generating responses for different applications. Given the dynamic nature of user preferences and the uncertainty of LLM response performance, it is crucial to design efficient online learning algorithms to identify optimal LLM responses (i.e., high-quality responses that also meet user preferences). Most existing online algorithms adopt a centralized approach and fail to leverage explicit user preferences for more efficient and personalized LLM response identification. In contrast, this paper introduces \textit{MACO} (\underline{M}ulti-\underline{A}gent \underline{C}onversational \underline{O}nline Learning for Adaptive LLM Response Identification): 1) The online LLM response identification process is accelerated by multiple local agents (such as smartphones), while enhancing data privacy; 2) A novel conversational mechanism is proposed to adaptively conduct conversations for soliciting user preferences (e.g., a preference for a humorous tone over a serious one in generated responses), so to minimize uncertainty in preference estimation. Our theoretical analysis demonstrates that \cadi\ is near-optimal regarding cumulative regret. Additionally, \cadi\ offers reduced communication costs and computational complexity by eliminating the traditional, computing-intensive ``G-optimal design" found in previous works. Extensive experiments with the open LLM \textit{Llama}, coupled with two different embedding models from Google and OpenAI for text vector representation, demonstrate that \cadi\ significantly outperforms the current state-of-the-art in online LLM response identification.
2501.01850
LCFed: An Efficient Clustered Federated Learning Framework for Heterogeneous Data
cs.LG cs.AI cs.DC
Clustered federated learning (CFL) addresses the performance challenges posed by data heterogeneity in federated learning (FL) by organizing edge devices with similar data distributions into clusters, enabling collaborative model training tailored to each group. However, existing CFL approaches strictly limit knowledge sharing to within clusters, lacking the integration of global knowledge with intra-cluster training, which leads to suboptimal performance. Moreover, traditional clustering methods incur significant computational overhead, especially as the number of edge devices increases. In this paper, we propose LCFed, an efficient CFL framework to combat these challenges. By leveraging model partitioning and adopting distinct aggregation strategies for each sub-model, LCFed effectively incorporates global knowledge into intra-cluster co-training, achieving optimal training performance. Additionally, LCFed customizes a computationally efficient model similarity measurement method based on low-rank models, enabling real-time cluster updates with minimal computational overhead. Extensive experiments show that LCFed outperforms state-of-the-art benchmarks in both test accuracy and clustering computational efficiency.
2501.01855
UAV-DETR: Efficient End-to-End Object Detection for Unmanned Aerial Vehicle Imagery
cs.CV
Unmanned aerial vehicle object detection (UAV-OD) has been widely used in various scenarios. However, most existing UAV-OD algorithms rely on manually designed components, which require extensive tuning. End-to-end models that do not depend on such manually designed components are mainly designed for natural images, which are less effective for UAV imagery. To address such challenges, this paper proposes an efficient detection transformer (DETR) framework tailored for UAV imagery, i.e., UAV-DETR. The framework includes a multi-scale feature fusion with frequency enhancement module, which captures both spatial and frequency information at different scales. In addition, a frequency-focused down-sampling module is presented to retain critical spatial details during down-sampling. A semantic alignment and calibration module is developed to align and fuse features from different fusion paths. Experimental results demonstrate the effectiveness and generalization of our approach across various UAV imagery datasets. On the VisDrone dataset, our method improves AP by 3.1\% and $\text{AP}_{50}$ by 4.2\% over the baseline. Similar enhancements are observed on the UAVVaste dataset. The project page: https://github.com/ValiantDiligent/UAV-DETR
2501.01859
Deposition Rates in Thermal Laser Epitaxy: Simulation and Experiment
cs.CE physics.comp-ph
The modeling of deposition rates in Thermal Laser Epitaxy (TLE) is essential for the accurate prediction of the evaporation process and for improved dynamic process control. We demonstrate excellent agreement between experimental data and a model based on a finite element simulation that describes the temperature distribution of an elemental source when irradiated with continuous wave laser radiation. The simulation strongly depends on the thermophysical constants of the material, data of which is lacking for many elements. Effective values for the parameters may be determined with precision by means of an unambiguous reference provided by the melting point of the material, which is directly observed during the experiments. TLE may therefore be used to study the high temperature thermophysical and optical properties of the elements.
2501.01864
Towards Hard and Soft Shadow Removal via Dual-Branch Separation Network and Vision Transformer
cs.CV
Image shadow removal is a crucial task in computer vision. In real-world scenes, shadows alter image color and brightness, posing challenges for perception and texture recognition. Traditional and deep learning methods often overlook the distinct needs for handling hard and soft shadows, thereby lacking detailed processing to specifically address each type of shadow in images.We propose a dual-path model that processes these shadows separately using specially designed loss functions to accomplish the hard and soft shadow removal. The model classifies shadow types and processes them through appropriate paths to produce shadow-free outputs, integrating a Vision Transformer with UNet++ for enhanced edge detail and feature fusion. Our model outperforms state-of-the-art methods and achieves 2.905 RMSE value on the ISTD dataset, which demonstrates greater effectiveness than typical single-path approaches.
2501.01872
Turning Logic Against Itself : Probing Model Defenses Through Contrastive Questions
cs.CL
Large language models, despite extensive alignment with human values and ethical principles, remain vulnerable to sophisticated jailbreak attacks that exploit their reasoning abilities. Existing safety measures often detect overt malicious intent but fail to address subtle, reasoning-driven vulnerabilities. In this work, we introduce POATE (Polar Opposite query generation, Adversarial Template construction, and Elaboration), a novel jailbreak technique that harnesses contrastive reasoning to provoke unethical responses. POATE crafts semantically opposing intents and integrates them with adversarial templates, steering models toward harmful outputs with remarkable subtlety. We conduct extensive evaluation across six diverse language model families of varying parameter sizes to demonstrate the robustness of the attack, achieving significantly higher attack success rates (~44%) compared to existing methods. To counter this, we propose Intent-Aware CoT and Reverse Thinking CoT, which decompose queries to detect malicious intent and reason in reverse to evaluate and reject harmful responses. These methods enhance reasoning robustness and strengthen the model's defense against adversarial exploits.
2501.01874
DFF: Decision-Focused Fine-tuning for Smarter Predict-then-Optimize with Limited Data
cs.LG
Decision-focused learning (DFL) offers an end-to-end approach to the predict-then-optimize (PO) framework by training predictive models directly on decision loss (DL), enhancing decision-making performance within PO contexts. However, the implementation of DFL poses distinct challenges. Primarily, DL can result in deviation from the physical significance of the predictions under limited data. Additionally, some predictive models are non-differentiable or black-box, which cannot be adjusted using gradient-based methods. To tackle the above challenges, we propose a novel framework, Decision-Focused Fine-tuning (DFF), which embeds the DFL module into the PO pipeline via a novel bias correction module. DFF is formulated as a constrained optimization problem that maintains the proximity of the DL-enhanced model to the original predictive model within a defined trust region. We theoretically prove that DFF strictly confines prediction bias within a predetermined upper bound, even with limited datasets, thereby substantially reducing prediction shifts caused by DL under limited data. Furthermore, the bias correction module can be integrated into diverse predictive models, enhancing adaptability to a broad range of PO tasks. Extensive evaluations on synthetic and real-world datasets, including network flow, portfolio optimization, and resource allocation problems with different predictive models, demonstrate that DFF not only improves decision performance but also adheres to fine-tuning constraints, showcasing robust adaptability across various scenarios.
2501.01876
Accuracy Can Lie: On the Impact of Surrogate Model in Configuration Tuning
cs.SE cs.AI
To ease the expensive measurements during configuration tuning, it is natural to build a surrogate model as the replacement of the system, and thereby the configuration performance can be cheaply evaluated. Yet, a stereotype therein is that the higher the model accuracy, the better the tuning result would be. This "accuracy is all" belief drives our research community to build more and more accurate models and criticize a tuner for the inaccuracy of the model used. However, this practice raises some previously unaddressed questions, e.g., Do those somewhat small accuracy improvements reported in existing work really matter much to the tuners? What role does model accuracy play in the impact of tuning quality? To answer those related questions, we conduct one of the largest-scale empirical studies to date-running over the period of 13 months 24*7-that covers 10 models, 17 tuners, and 29 systems from the existing works while under four different commonly used metrics, leading to 13,612 cases of investigation. Surprisingly, our key findings reveal that the accuracy can lie: there are a considerable number of cases where higher accuracy actually leads to no improvement in the tuning outcomes (up to 58% cases under certain setting), or even worse, it can degrade the tuning quality (up to 24% cases under certain setting). We also discover that the chosen models in most proposed tuners are sub-optimal and that the required % of accuracy change to significantly improve tuning quality varies according to the range of model accuracy. Deriving from the fitness landscape analysis, we provide in-depth discussions of the rationale behind, offering several lessons learned as well as insights for future opportunities. Most importantly, this work poses a clear message to the community: we should take one step back from the natural "accuracy is all" belief for model-based configuration tuning.
2501.01877
ANTHROPOS-V: benchmarking the novel task of Crowd Volume Estimation
cs.CV
We introduce the novel task of Crowd Volume Estimation (CVE), defined as the process of estimating the collective body volume of crowds using only RGB images. Besides event management and public safety, CVE can be instrumental in approximating body weight, unlocking weight sensitive applications such as infrastructure stress assessment, and assuring even weight balance. We propose the first benchmark for CVE, comprising ANTHROPOS-V, a synthetic photorealistic video dataset featuring crowds in diverse urban environments. Its annotations include each person's volume, SMPL shape parameters, and keypoints. Also, we explore metrics pertinent to CVE, define baseline models adapted from Human Mesh Recovery and Crowd Counting domains, and propose a CVE specific methodology that surpasses baselines. Although synthetic, the weights and heights of individuals are aligned with the real-world population distribution across genders, and they transfer to the downstream task of CVE from real images. Benchmark and code are available at github.com/colloroneluca/Crowd-Volume-Estimation.
2501.01880
Long Context vs. RAG for LLMs: An Evaluation and Revisits
cs.CL
Extending context windows (i.e., Long Context, LC) and using retrievers to selectively access relevant information (i.e., Retrieval-Augmented Generation, RAG) are the two main strategies to enable LLMs to incorporate extremely long external contexts. This paper revisits recent studies on this topic, highlighting their key insights and discrepancies. We then provide a more comprehensive evaluation by filtering out questions answerable without external context, identifying the most effective retrieval methods, and expanding the datasets. We show that LC generally outperforms RAG in question-answering benchmarks, especially for Wikipedia-based questions. Summarization-based retrieval performs comparably to LC, while chunk-based retrieval lags behind. However, RAG has advantages in dialogue-based and general question queries. These insights underscore the trade-offs between RAG and LC strategies, offering guidance for future optimization of LLMs with external knowledge sources. We also provide an in-depth discussion on this topic, highlighting the overlooked importance of context relevance in existing studies.
2501.01884
Telegram as a Battlefield: Kremlin-related Communications during the Russia-Ukraine Conflict
cs.SI cs.CY cs.HC
Telegram emerged as a crucial platform for both parties during the conflict between Russia and Ukraine. Per its minimal policies for content moderation, Pro-Kremlin narratives and potential misinformation were spread on Telegram, while anti-Kremlin narratives with related content were also propagated, such as war footage, troop movements, maps of bomb shelters, and air raid warnings. This paper presents a dataset of posts from both pro-Kremlin and anti-Kremlin Telegram channels, collected over a period spanning a year before and a year after the Russian invasion. The dataset comprises 404 pro-Kremlin channels with 4,109,645 posts and 114 anti-Kremlin channels with 1,117,768 posts. We provide details on the data collection process, processing methods, and dataset characterization. Lastly, we discuss the potential research opportunities this dataset may enable researchers across various disciplines.
2501.01886
Evaluating Scenario-based Decision-making for Interactive Autonomous Driving Using Rational Criteria: A Survey
cs.RO cs.AI cs.SY eess.SY
Autonomous vehicles (AVs) can significantly promote the advances in road transport mobility in terms of safety, reliability, and decarbonization. However, ensuring safety and efficiency in interactive during within dynamic and diverse environments is still a primary barrier to large-scale AV adoption. In recent years, deep reinforcement learning (DRL) has emerged as an advanced AI-based approach, enabling AVs to learn decision-making strategies adaptively from data and interactions. DRL strategies are better suited than traditional rule-based methods for handling complex, dynamic, and unpredictable driving environments due to their adaptivity. However, varying driving scenarios present distinct challenges, such as avoiding obstacles on highways and reaching specific exits at intersections, requiring different scenario-specific decision-making algorithms. Many DRL algorithms have been proposed in interactive decision-making. However, a rationale review of these DRL algorithms across various scenarios is lacking. Therefore, a comprehensive evaluation is essential to assess these algorithms from multiple perspectives, including those of vehicle users and vehicle manufacturers. This survey reviews the application of DRL algorithms in autonomous driving across typical scenarios, summarizing road features and recent advancements. The scenarios include highways, on-ramp merging, roundabouts, and unsignalized intersections. Furthermore, DRL-based algorithms are evaluated based on five rationale criteria: driving safety, driving efficiency, training efficiency, unselfishness, and interpretability (DDTUI). Each criterion of DDTUI is specifically analyzed in relation to the reviewed algorithms. Finally, the challenges for future DRL-based decision-making algorithms are summarized.
2501.01889
Exploring Equality: An Investigation into Custom Loss Functions for Fairness Definitions
cs.LG cs.CY
This paper explores the complex tradeoffs between various fairness metrics such as equalized odds, disparate impact, and equal opportunity and predictive accuracy within COMPAS by building neural networks trained with custom loss functions optimized to specific fairness criteria. This paper creates the first fairness-driven implementation of the novel Group Accuracy Parity (GAP) framework, as theoretically proposed by Gupta et al. (2024), and applies it to COMPAS. To operationalize and accurately compare the fairness of COMPAS models optimized to differing fairness ideals, this paper develops and proposes a combinatory analytical procedure that incorporates Pareto front and multivariate analysis, leveraging data visualizations such as violin graphs. This paper concludes that GAP achieves an enhanced equilibrium between fairness and accuracy compared to COMPAS's current nationwide implementation and alternative implementations of COMPAS optimized to more traditional fairness definitions. While this paper's algorithmic improvements of COMPAS significantly augment its fairness, external biases undermine the fairness of its implementation. Practices such as predictive policing and issues such as the lack of transparency regarding COMPAS's internal workings have contributed to the algorithm's historical injustice. In conjunction with developments regarding COMPAS's predictive methodology, legal and institutional changes must happen for COMPAS's just deployment.
2501.01892
QuArch: A Question-Answering Dataset for AI Agents in Computer Architecture
cs.AR cs.AI cs.LG
We introduce QuArch, a dataset of 1500 human-validated question-answer pairs designed to evaluate and enhance language models' understanding of computer architecture. The dataset covers areas including processor design, memory systems, and performance optimization. Our analysis highlights a significant performance gap: the best closed-source model achieves 84% accuracy, while the top small open-source model reaches 72%. We observe notable struggles in memory systems, interconnection networks, and benchmarking. Fine-tuning with QuArch improves small model accuracy by up to 8%, establishing a foundation for advancing AI-driven computer architecture research. The dataset and leaderboard are at https://harvard-edge.github.io/QuArch/.
2501.01895
EnerVerse: Envisioning Embodied Future Space for Robotics Manipulation
cs.RO cs.CV cs.LG
We introduce EnerVerse, a generative robotics foundation model that constructs and interprets embodied spaces. EnerVerse employs an autoregressive video diffusion framework to predict future embodied spaces from instructions, enhanced by a sparse context memory for long-term reasoning. To model the 3D robotics world, we propose Free Anchor Views (FAVs), a multi-view video representation offering flexible, task-adaptive perspectives to address challenges like motion ambiguity and environmental constraints. Additionally, we present EnerVerse-D, a data engine pipeline combining the generative model with 4D Gaussian Splatting, forming a self-reinforcing data loop to reduce the sim-to-real gap. Leveraging these innovations, EnerVerse translates 4D world representations into physical actions via a policy head (EnerVerse-A), enabling robots to execute task instructions. EnerVerse-A achieves state-of-the-art performance in both simulation and real-world settings.
2501.01904
Virgo: A Preliminary Exploration on Reproducing o1-like MLLM
cs.CV cs.AI
Recently, slow-thinking reasoning systems, built upon large language models (LLMs), have garnered widespread attention by scaling the thinking time during inference. There is also growing interest in adapting this capability to multimodal large language models (MLLMs). Given that MLLMs handle more complex data semantics across different modalities, it is intuitively more challenging to implement multimodal slow-thinking systems. To address this issue, in this paper, we explore a straightforward approach by fine-tuning a capable MLLM with a small amount of textual long-form thought data, resulting in a multimodal slow-thinking system, Virgo (Visual reasoning with long thought). We find that these long-form reasoning processes, expressed in natural language, can be effectively transferred to MLLMs. Moreover, it seems that such textual reasoning data can be even more effective than visual reasoning data in eliciting the slow-thinking capacities of MLLMs. While this work is preliminary, it demonstrates that slow-thinking capacities are fundamentally associated with the language model component, which can be transferred across modalities or domains. This finding can be leveraged to guide the development of more powerful slow-thinking reasoning systems. We release our resources at https://github.com/RUCAIBox/Virgo.
2501.01905
Alleviating Overfitting in Transformation-Interaction-Rational Symbolic Regression with Multi-Objective Optimization
cs.LG
The Transformation-Interaction-Rational is a representation for symbolic regression that limits the search space of functions to the ratio of two nonlinear functions each one defined as the linear regression of transformed variables. This representation has the main objective to bias the search towards simpler expressions while keeping the approximation power of standard approaches. The performance of using Genetic Programming with this representation was substantially better than with its predecessor (Interaction-Transformation) and ranked close to the state-of-the-art on a contemporary Symbolic Regression benchmark. On a closer look at these results, we observed that the performance could be further improved with an additional selective pressure for smaller expressions when the dataset contains just a few data points. The introduction of a penalization term applied to the fitness measure improved the results on these smaller datasets. One problem with this approach is that it introduces two additional hyperparameters: i) a criteria to when the penalization should be activated and, ii) the amount of penalization to the fitness function. In this paper, we extend Transformation-Interaction-Rational to support multi-objective optimization, specifically the NSGA-II algorithm, and apply that to the same benchmark. A detailed analysis of the results show that the use of multi-objective optimization benefits the overall performance on a subset of the benchmarks while keeping the results similar to the single-objective approach on the remainder of the datasets. Specifically to the small datasets, we observe a small (and statistically insignificant) improvement of the results suggesting that further strategies must be explored.
2501.01908
Detecting and Mitigating Adversarial Attacks on Deep Learning-Based MRI Reconstruction Without Any Retraining
cs.CV cs.LG eess.IV physics.med-ph
Deep learning (DL) methods, especially those based on physics-driven DL, have become the state-of-the-art for reconstructing sub-sampled magnetic resonance imaging (MRI) data. However, studies have shown that these methods are susceptible to small adversarial input perturbations, or attacks, resulting in major distortions in the output images. Various strategies have been proposed to reduce the effects of these attacks, but they require retraining and may lower reconstruction quality for non-perturbed/clean inputs. In this work, we propose a novel approach for detecting and mitigating adversarial attacks on MRI reconstruction models without any retraining. Our detection strategy is based on the idea of cyclic measurement consistency. The output of the model is mapped to another set of MRI measurements for a different sub-sampling pattern, and this synthesized data is reconstructed with the same model. Intuitively, without an attack, the second reconstruction is expected to be consistent with the first, while with an attack, disruptions are present. Subsequently, this idea is extended to devise a novel objective function, which is minimized within a small ball around the attack input for mitigation. Experimental results show that our method substantially reduces the impact of adversarial perturbations across different datasets, attack types/strengths and PD-DL networks, and qualitatively and quantitatively outperforms conventional mitigation methods that involve retraining.
2501.01912
Exoplanet Detection via Differentiable Rendering
astro-ph.EP astro-ph.IM cs.CV eess.IV
Direct imaging of exoplanets is crucial for advancing our understanding of planetary systems beyond our solar system, but it faces significant challenges due to the high contrast between host stars and their planets. Wavefront aberrations introduce speckles in the telescope science images, which are patterns of diffracted starlight that can mimic the appearance of planets, complicating the detection of faint exoplanet signals. Traditional post-processing methods, operating primarily in the image intensity domain, do not integrate wavefront sensing data. These data, measured mainly for adaptive optics corrections, have been overlooked as a potential resource for post-processing, partly due to the challenge of the evolving nature of wavefront aberrations. In this paper, we present a differentiable rendering approach that leverages these wavefront sensing data to improve exoplanet detection. Our differentiable renderer models wave-based light propagation through a coronagraphic telescope system, allowing gradient-based optimization to significantly improve starlight subtraction and increase sensitivity to faint exoplanets. Simulation experiments based on the James Webb Space Telescope configuration demonstrate the effectiveness of our approach, achieving substantial improvements in contrast and planet detection limits. Our results showcase how the computational advancements enabled by differentiable rendering can revitalize previously underexploited wavefront data, opening new avenues for enhancing exoplanet imaging and characterization.
2501.01913
Mingling with the Good to Backdoor Federated Learning
cs.CR cs.AI cs.DC
Federated learning (FL) is a decentralized machine learning technique that allows multiple entities to jointly train a model while preserving dataset privacy. However, its distributed nature has raised various security concerns, which have been addressed by increasingly sophisticated defenses. These protections utilize a range of data sources and metrics to, for example, filter out malicious model updates, ensuring that the impact of attacks is minimized or eliminated. This paper explores the feasibility of designing a generic attack method capable of installing backdoors in FL while evading a diverse array of defenses. Specifically, we focus on an attacker strategy called MIGO, which aims to produce model updates that subtly blend with legitimate ones. The resulting effect is a gradual integration of a backdoor into the global model, often ensuring its persistence long after the attack concludes, while generating enough ambiguity to hinder the effectiveness of defenses. MIGO was employed to implant three types of backdoors across five datasets and different model architectures. The results demonstrate the significant threat posed by these backdoors, as MIGO consistently achieved exceptionally high backdoor accuracy (exceeding 90%) while maintaining the utility of the main task. Moreover, MIGO exhibited strong evasion capabilities against ten defenses, including several state-of-the-art methods. When compared to four other attack strategies, MIGO consistently outperformed them across most configurations. Notably, even in extreme scenarios where the attacker controls just 0.1% of the clients, the results indicate that successful backdoor insertion is possible if the attacker can persist for a sufficient number of rounds.
2501.01915
Social Processes: Probabilistic Meta-learning for Adaptive Multiparty Interaction Forecasting
cs.LG
Adaptively forecasting human behavior in social settings is an important step toward achieving Artificial General Intelligence. Most existing research in social forecasting has focused either on unfocused interactions, such as pedestrian trajectory prediction, or on monadic and dyadic behavior forecasting. In contrast, social psychology emphasizes the importance of group interactions for understanding complex social dynamics. This creates a gap that we address in this paper: forecasting social interactions at the group (conversation) level. Additionally, it is important for a forecasting model to be able to adapt to groups unseen at train time, as even the same individual behaves differently across different groups. This highlights the need for a forecasting model to explicitly account for each group's unique dynamics. To achieve this, we adopt a meta-learning approach to human behavior forecasting, treating every group as a separate meta-learning task. As a result, our method conditions its predictions on the specific behaviors within the group, leading to generalization to unseen groups. Specifically, we introduce Social Process (SP) models, which predict a distribution over future multimodal cues jointly for all group members based on their preceding low-level multimodal cues, while incorporating other past sequences of the same group's interactions. In this work we also analyze the generalization capabilities of SP models in both their outputs and latent spaces through the use of realistic synthetic datasets.
2501.01924
Transformer-Driven Inverse Problem Transform for Fast Blind Hyperspectral Image Dehazing
cs.CV eess.IV
Hyperspectral dehazing (HyDHZ) has become a crucial signal processing technology to facilitate the subsequent identification and classification tasks, as the airborne visible/infrared imaging spectrometer (AVIRIS) data portal reports a massive portion of haze-corrupted areas in typical hyperspectral remote sensing images. The idea of inverse problem transform (IPT) has been proposed in recent remote sensing literature in order to reformulate a hardly tractable inverse problem (e.g., HyDHZ) into a relatively simple one. Considering the emerging spectral super-resolution (SSR) technique, which spectrally upsamples multispectral data to hyperspectral data, we aim to solve the challenging HyDHZ problem by reformulating it as an SSR problem. Roughly speaking, the proposed algorithm first automatically selects some uncorrupted/informative spectral bands, from which SSR is applied to spectrally upsample the selected bands in the feature space, thereby obtaining a clean hyperspectral image (HSI). The clean HSI is then further refined by a deep transformer network to obtain the final dehazed HSI, where a global attention mechanism is designed to capture nonlocal information. There are very few HyDHZ works in existing literature, and this article introduces the powerful spatial-spectral transformer into HyDHZ for the first time. Remarkably, the proposed transformer-driven IPT-based HyDHZ (T2HyDHZ) is a blind algorithm without requiring the user to manually select the corrupted region. Extensive experiments demonstrate the superiority of T2HyDHZ with less color distortion.
2501.01926
Mitigating Hallucination for Large Vision Language Model by Inter-Modality Correlation Calibration Decoding
cs.CV cs.AI
Large vision-language models (LVLMs) have shown remarkable capabilities in visual-language understanding for downstream multi-modal tasks. Despite their success, LVLMs still suffer from generating hallucinations in complex generation tasks, leading to inconsistencies between visual inputs and generated content. To address this issue, some approaches have introduced inference-time interventions, such as contrastive decoding and attention rectification, to reduce overreliance on language priors. However, these approaches overlook hallucinations stemming from spurious inter-modality correlations. In this paper, we propose an Inter-Modality Correlation Calibration Decoding (IMCCD) method to mitigate hallucinations in LVLMs in a training-free manner. In this method, we design a Cross-Modal Value-Enhanced Decoding(CMVED) module to alleviate hallucination by a novel contrastive decoding mechanism. During the estimation of distorted distribution, CMVED masks the value vectors associated with significant cross-modal attention weights, which address both uni-modality overreliance and misleading inter-modality correlations. Additionally, a Content-Driven Attention Refinement(CDAR) module refines cross-modal attention weights, guiding LVLMs to focus on important visual content. Experimental results on diverse hallucination benchmarks validate the superiority of our method over existing state-of-the-art techniques in reducing hallucinations in LVLM text generation. Our code will be available at https://github.com/lijm48/IMCCD.
2501.01929
Compressed sensing for inverse problems II: applications to deconvolution, source recovery, and MRI
math.FA cs.IT math.IT math.OC
This paper extends the sample complexity theory for ill-posed inverse problems developed in a recent work by the authors [`Compressed sensing for inverse problems and the sample complexity of the sparse Radon transform', J. Eur. Math. Soc., to appear], which was originally focused on the sparse Radon transform. We demonstrate that the underlying abstract framework, based on infinite-dimensional compressed sensing and generalized sampling techniques, can effectively handle a variety of practical applications. Specifically, we analyze three case studies: (1) The reconstruction of a sparse signal from a finite number of pointwise blurred samples; (2) The recovery of the (sparse) source term of an elliptic partial differential equation from finite samples of the solution; and (3) A moderately ill-posed variation of the classical sensing problem of recovering a wavelet-sparse signal from finite Fourier samples, motivated by magnetic resonance imaging. For each application, we establish rigorous recovery guarantees by verifying the key theoretical requirements, including quasi-diagonalization and coherence bounds. Our analysis reveals that careful consideration of balancing properties and optimized sampling strategies can lead to improved reconstruction performance. The results provide a unified theoretical foundation for compressed sensing approaches to inverse problems while yielding practical insights for specific applications.
2501.01930
GoBERT: Gene Ontology Graph Informed BERT for Universal Gene Function Prediction
cs.LG
Exploring the functions of genes and gene products is crucial to a wide range of fields, including medical research, evolutionary biology, and environmental science. However, discovering new functions largely relies on expensive and exhaustive wet lab experiments. Existing methods of automatic function annotation or prediction mainly focus on protein function prediction with sequence, 3D-structures or protein family information. In this study, we propose to tackle the gene function prediction problem by exploring Gene Ontology graph and annotation with BERT (GoBERT) to decipher the underlying relationships among gene functions. Our proposed novel function prediction task utilizes existing functions as inputs and generalizes the function prediction to gene and gene products. Specifically, two pre-train tasks are designed to jointly train GoBERT to capture both explicit and implicit relations of functions. Neighborhood prediction is a self-supervised multi-label classification task that captures the explicit function relations. Specified masking and recovering task helps GoBERT in finding implicit patterns among functions. The pre-trained GoBERT possess the ability to predict novel functions for various gene and gene products based on known functional annotations. Extensive experiments, biological case studies, and ablation studies are conducted to demonstrate the superiority of our proposed GoBERT.
2501.01932
Bridging Classification and Segmentation in Osteosarcoma Assessment via Foundation and Discrete Diffusion Models
cs.CV
Osteosarcoma, the most common primary bone cancer, often requires accurate necrosis assessment from whole slide images (WSIs) for effective treatment planning and prognosis. However, manual assessments are subjective and prone to variability. In response, we introduce FDDM, a novel framework bridging the gap between patch classification and region-based segmentation. FDDM operates in two stages: patch-based classification, followed by region-based refinement, enabling cross-patch information intergation. Leveraging a newly curated dataset of osteosarcoma images, FDDM demonstrates superior segmentation performance, achieving up to a 10% improvement mIOU and a 32.12% enhancement in necrosis rate estimation over state-of-the-art methods. This framework sets a new benchmark in osteosarcoma assessment, highlighting the potential of foundation models and diffusion-based refinements in complex medical imaging tasks.
2501.01933
Abstractive Text Summarization for Contemporary Sanskrit Prose: Issues and Challenges
cs.CL cs.AI
This thesis presents Abstractive Text Summarization models for contemporary Sanskrit prose. The first chapter, titled Introduction, presents the motivation behind this work, the research questions, and the conceptual framework. Sanskrit is a low-resource inflectional language. The key research question that this thesis investigates is what the challenges in developing an abstractive TS for Sanskrit. To answer the key research questions, sub-questions based on four different themes have been posed in this work. The second chapter, Literature Review, surveys the previous works done. The third chapter, data preparation, answers the remaining three questions from the third theme. It reports the data collection and preprocessing challenges for both language model and summarization model trainings. The fourth chapter reports the training and inference of models and the results obtained therein. This research has initiated a pipeline for Sanskrit abstractive text summarization and has reported the challenges faced at every stage of the development. The research questions based on every theme have been answered to answer the key research question.
2501.01934
Fusion DeepONet: A Data-Efficient Neural Operator for Geometry-Dependent Hypersonic Flows on Arbitrary Grids
cs.LG physics.flu-dyn
Designing re-entry vehicles requires accurate predictions of hypersonic flow around their geometry. Rapid prediction of such flows can revolutionize vehicle design, particularly for morphing geometries. We evaluate advanced neural operator models such as Deep Operator Networks (DeepONet), parameter-conditioned U-Net, Fourier Neural Operator (FNO), and MeshGraphNet, with the objective of addressing the challenge of learning geometry-dependent hypersonic flow fields with limited data. Specifically, we compare the performance of these models for two grid types: uniform Cartesian and irregular grids. To train these models, we use 36 unique elliptic geometries for generating high-fidelity simulations with a high-order entropy-stable DGSEM solver, emphasizing the challenge of working with a scarce dataset. We evaluate and compare the four operator-based models for their efficacy in predicting hypersonic flow field around the elliptic body. Moreover, we develop a novel framework, called Fusion DeepONet, which leverages neural field concepts and generalizes effectively across varying geometries. Despite the scarcity of training data, Fusion DeepONet achieves performance comparable to parameter-conditioned U-Net on uniform grids while it outperforms MeshGraphNet and vanilla DeepONet on irregular, arbitrary grids. Fusion DeepONet requires significantly fewer trainable parameters as compared to U-Net, MeshGraphNet, and FNO, making it computationally efficient. We also analyze the basis functions of the Fusion DeepONet model using Singular Value Decomposition. This analysis reveals that Fusion DeepONet generalizes effectively to unseen solutions and adapts to varying geometries and grid points, demonstrating its robustness in scenarios with limited training data.
2501.01936
Improving Transducer-Based Spoken Language Understanding with Self-Conditioned CTC and Knowledge Transfer
cs.LG
In this paper, we propose to improve end-to-end (E2E) spoken language understand (SLU) in an RNN transducer model (RNN-T) by incorporating a joint self-conditioned CTC automatic speech recognition (ASR) objective. Our proposed model is akin to an E2E differentiable cascaded model which performs ASR and SLU sequentially and we ensure that the SLU task is conditioned on the ASR task by having CTC self conditioning. This novel joint modeling of ASR and SLU improves SLU performance significantly over just using SLU optimization. We further improve the performance by aligning the acoustic embeddings of this model with the semantically richer BERT model. Our proposed knowledge transfer strategy makes use of a bag-of-entity prediction layer on the aligned embeddings and the output of this is used to condition the RNN-T based SLU decoding. These techniques show significant improvement over several strong baselines and can perform at par with large models like Whisper with significantly fewer parameters.
2501.01945
Cold-Start Recommendation towards the Era of Large Language Models (LLMs): A Comprehensive Survey and Roadmap
cs.IR cs.AI
Cold-start problem is one of the long-standing challenges in recommender systems, focusing on accurately modeling new or interaction-limited users or items to provide better recommendations. Due to the diversification of internet platforms and the exponential growth of users and items, the importance of cold-start recommendation (CSR) is becoming increasingly evident. At the same time, large language models (LLMs) have achieved tremendous success and possess strong capabilities in modeling user and item information, providing new potential for cold-start recommendations. However, the research community on CSR still lacks a comprehensive review and reflection in this field. Based on this, in this paper, we stand in the context of the era of large language models and provide a comprehensive review and discussion on the roadmap, related literature, and future directions of CSR. Specifically, we have conducted an exploration of the development path of how existing CSR utilizes information, from content features, graph relations, and domain information, to the world knowledge possessed by large language models, aiming to provide new insights for both the research and industrial communities on CSR. Related resources of cold-start recommendations are collected and continuously updated for the community in https://github.com/YuanchenBei/Awesome-Cold-Start-Recommendation.
2501.01949
VideoLifter: Lifting Videos to 3D with Fast Hierarchical Stereo Alignment
cs.CV
Efficiently reconstructing accurate 3D models from monocular video is a key challenge in computer vision, critical for advancing applications in virtual reality, robotics, and scene understanding. Existing approaches typically require pre-computed camera parameters and frame-by-frame reconstruction pipelines, which are prone to error accumulation and entail significant computational overhead. To address these limitations, we introduce VideoLifter, a novel framework that leverages geometric priors from a learnable model to incrementally optimize a globally sparse to dense 3D representation directly from video sequences. VideoLifter segments the video sequence into local windows, where it matches and registers frames, constructs consistent fragments, and aligns them hierarchically to produce a unified 3D model. By tracking and propagating sparse point correspondences across frames and fragments, VideoLifter incrementally refines camera poses and 3D structure, minimizing reprojection error for improved accuracy and robustness. This approach significantly accelerates the reconstruction process, reducing training time by over 82% while surpassing current state-of-the-art methods in visual fidelity and computational efficiency.
2501.01950
MADGEN: Mass-Spec attends to De Novo Molecular generation
cs.LG cs.AI
The annotation (assigning structural chemical identities) of MS/MS spectra remains a significant challenge due to the enormous molecular diversity in biological samples and the limited scope of reference databases. Currently, the vast majority of spectral measurements remain in the "dark chemical space" without structural annotations. To improve annotation, we propose MADGEN (Mass-spec Attends to De Novo Molecular GENeration), a scaffold-based method for de novo molecular structure generation guided by mass spectrometry data. MADGEN operates in two stages: scaffold retrieval and spectra-conditioned molecular generation starting with the scaffold. In the first stage, given an MS/MS spectrum, we formulate scaffold retrieval as a ranking problem and employ contrastive learning to align mass spectra with candidate molecular scaffolds. In the second stage, starting from the retrieved scaffold, we employ the MS/MS spectrum to guide an attention-based generative model to generate the final molecule. Our approach constrains the molecular generation search space, reducing its complexity and improving generation accuracy. We evaluate MADGEN on three datasets (NIST23, CANOPUS, and MassSpecGym) and evaluate MADGEN's performance with a predictive scaffold retriever and with an oracle retriever. We demonstrate the effectiveness of using attention to integrate spectral information throughout the generation process to achieve strong results with the oracle retriever.
2501.01951
MixGCN: Scalable GCN Training by Mixture of Parallelism and Mixture of Accelerators
cs.LG cs.AI
Graph convolutional networks (GCNs) have demonstrated superiority in graph-based learning tasks. However, training GCNs on full graphs is particularly challenging, due to the following two challenges: (1) the associated feature tensors can easily explode the memory and block the communication bandwidth of modern accelerators, and (2) the computation workflow in training GCNs alternates between sparse and dense matrix operations, complicating the efficient utilization of computational resources. Existing solutions for scalable distributed full-graph GCN training mostly adopt partition parallelism, which is unsatisfactory as they only partially address the first challenge while incurring scaled-out communication volume. To this end, we propose MixGCN aiming to simultaneously address both the aforementioned challenges towards GCN training. To tackle the first challenge, MixGCN integrates mixture of parallelism. Both theoretical and empirical analysis verify its constant communication volumes and enhanced balanced workload; For handling the second challenge, we consider mixture of accelerators (i.e., sparse and dense accelerators) with a dedicated accelerator for GCN training and a fine-grain pipeline. Extensive experiments show that MixGCN achieves boosted training efficiency and scalability.
2501.01956
Metadata Conditioning Accelerates Language Model Pre-training
cs.CL
The vast diversity of styles, domains, and quality levels present in language model pre-training corpora is essential in developing general model capabilities, but efficiently learning and deploying the correct behaviors exemplified in each of these heterogeneous data sources is challenging. To address this, we propose a new method, termed Metadata Conditioning then Cooldown (MeCo), to incorporate additional learning cues during pre-training. MeCo first provides metadata (e.g., URLs like en.wikipedia.org) alongside the text during training and later uses a cooldown phase with only the standard text, thereby enabling the model to function normally even without metadata. MeCo significantly accelerates pre-training across different model scales (600M to 8B parameters) and training sources (C4, RefinedWeb, and DCLM). For instance, a 1.6B language model trained with MeCo matches the downstream task performance of standard pre-training while using 33% less data. Additionally, MeCo enables us to steer language models by conditioning the inference prompt on either real or fabricated metadata that encodes the desired properties of the output: for example, prepending wikipedia.org to reduce harmful generations or factquizmaster.com (fabricated) to improve common knowledge task performance. We also demonstrate that MeCo is compatible with different types of metadata, such as model-generated topics. MeCo is remarkably simple, adds no computational overhead, and demonstrates promise in producing more capable and steerable language models.
2501.01957
VITA-1.5: Towards GPT-4o Level Real-Time Vision and Speech Interaction
cs.CV cs.SD eess.AS
Recent Multimodal Large Language Models (MLLMs) have typically focused on integrating visual and textual modalities, with less emphasis placed on the role of speech in enhancing interaction. However, speech plays a crucial role in multimodal dialogue systems, and implementing high-performance in both vision and speech tasks remains a significant challenge due to the fundamental modality differences. In this paper, we propose a carefully designed multi-stage training methodology that progressively trains LLM to understand both visual and speech information, ultimately enabling fluent vision and speech interaction. Our approach not only preserves strong vision-language capacity, but also enables efficient speech-to-speech dialogue capabilities without separate ASR and TTS modules, significantly accelerating multimodal end-to-end response speed. By comparing our method against state-of-the-art counterparts across benchmarks for image, video, and speech tasks, we demonstrate that our model is equipped with both strong visual and speech capabilities, making near real-time vision and speech interaction.
2501.01959
STEAM-EEG: Spatiotemporal EEG Analysis with Markov Transfer Fields and Attentive CNNs
cs.CV cs.AI cs.CE
Electroencephalogram (EEG) signals play a pivotal role in biomedical research and clinical applications, including epilepsy diagnosis, sleep disorder analysis, and brain-computer interfaces. However, the effective analysis and interpretation of these complex signals often present significant challenges. This paper presents a novel approach that integrates computer graphics techniques with biological signal pattern recognition, specifically using Markov Transfer Fields (MTFs) for EEG time series imaging. The proposed framework (STEAM-EEG) employs the capabilities of MTFs to capture the spatiotemporal dynamics of EEG signals, transforming them into visually informative images. These images are then rendered, visualised, and modelled using state-of-the-art computer graphics techniques, thereby facilitating enhanced data exploration, pattern recognition, and decision-making. The code could be accessed from GitHub.
2501.01960
GAF-FusionNet: Multimodal ECG Analysis via Gramian Angular Fields and Split Attention
cs.CV cs.AI cs.GR cs.LG
Electrocardiogram (ECG) analysis plays a crucial role in diagnosing cardiovascular diseases, but accurate interpretation of these complex signals remains challenging. This paper introduces a novel multimodal framework(GAF-FusionNet) for ECG classification that integrates time-series analysis with image-based representation using Gramian Angular Fields (GAF). Our approach employs a dual-layer cross-channel split attention module to adaptively fuse temporal and spatial features, enabling nuanced integration of complementary information. We evaluate GAF-FusionNet on three diverse ECG datasets: ECG200, ECG5000, and the MIT-BIH Arrhythmia Database. Results demonstrate significant improvements over state-of-the-art methods, with our model achieving 94.5\%, 96.9\%, and 99.6\% accuracy on the respective datasets. Our code will soon be available at https://github.com/Cross-Innovation-Lab/GAF-FusionNet.git.
2501.01963
Statistical learning does not always entail knowledge
cs.LG cs.AI cs.IT math.IT math.PR math.ST stat.ML stat.TH
In this paper, we study learning and knowledge acquisition (LKA) of an agent about a proposition that is either true or false. We use a Bayesian approach, where the agent receives data to update his beliefs about the proposition according to a posterior distribution. The LKA is formulated in terms of active information, with data representing external or exogenous information that modifies the agent's beliefs. It is assumed that data provide details about a number of features that are relevant to the proposition. We show that this leads to a Gibbs distribution posterior, which is in maximum entropy relative to the prior, conditioned on the side constraints that the data provide in terms of the features. We demonstrate that full learning is sometimes not possible and full knowledge acquisition is never possible when the number of extracted features is too small. We also distinguish between primary learning (receiving data about features of relevance for the proposition) and secondary learning (receiving data about the learning of another agent). We argue that this type of secondary learning does not represent true knowledge acquisition. Our results have implications for statistical learning algorithms, and we claim that such algorithms do not always generate true knowledge. The theory is illustrated with several examples.
2501.01969
Optimal bounds for dissatisfaction in perpetual voting
cs.GT cs.AI cs.LG
In perpetual voting, multiple decisions are made at different moments in time. Taking the history of previous decisions into account allows us to satisfy properties such as proportionality over periods of time. In this paper, we consider the following question: is there a perpetual approval voting method that guarantees that no voter is dissatisfied too many times? We identify a sufficient condition on voter behavior -- which we call 'bounded conflicts' condition -- under which a sublinear growth of dissatisfaction is possible. We provide a tight upper bound on the growth of dissatisfaction under bounded conflicts, using techniques from Kolmogorov complexity. We also observe that the approval voting with binary choices mimics the machine learning setting of prediction with expert advice. This allows us to present a voting method with sublinear guarantees on dissatisfaction under bounded conflicts, based on the standard techniques from prediction with expert advice.
2501.01973
INFELM: In-depth Fairness Evaluation of Large Text-To-Image Models
cs.CV cs.AI cs.CY
The rapid development of large language models (LLMs) and large vision models (LVMs) have propelled the evolution of multi-modal AI systems, which have demonstrated the remarkable potential for industrial applications by emulating human-like cognition. However, they also pose significant ethical challenges, including amplifying harmful content and reinforcing societal biases. For instance, biases in some industrial image generation models highlighted the urgent need for robust fairness assessments. Most existing evaluation frameworks focus on the comprehensiveness of various aspects of the models, but they exhibit critical limitations, including insufficient attention to content generation alignment and social bias-sensitive domains. More importantly, their reliance on pixel-detection techniques is prone to inaccuracies. To address these issues, this paper presents INFELM, an in-depth fairness evaluation on widely-used text-to-image models. Our key contributions are: (1) an advanced skintone classifier incorporating facial topology and refined skin pixel representation to enhance classification precision by at least 16.04%, (2) a bias-sensitive content alignment measurement for understanding societal impacts, (3) a generalizable representation bias evaluation for diverse demographic groups, and (4) extensive experiments analyzing large-scale text-to-image model outputs across six social-bias-sensitive domains. We find that existing models in the study generally do not meet the empirical fairness criteria, and representation bias is generally more pronounced than alignment errors. INFELM establishes a robust benchmark for fairness assessment, supporting the development of multi-modal AI systems that align with ethical and human-centric principles.
2501.01974
Hawkes based Representation Learning for Reasoning over Scale-free Community-structured Temporal Knowledge Graphs
cs.SI cs.AI cs.LG
Temporal knowledge graph (TKG) reasoning has become a hot topic due to its great value in many practical tasks. The key to TKG reasoning is modeling the structural information and evolutional patterns of the TKGs. While great efforts have been devoted to TKG reasoning, the structural and evolutional characteristics of real-world networks have not been considered. In the aspect of structure, real-world networks usually exhibit clear community structure and scale-free (long-tailed distribution) properties. In the aspect of evolution, the impact of an event decays with the time elapsing. In this paper, we propose a novel TKG reasoning model called Hawkes process-based Evolutional Representation Learning Network (HERLN), which learns structural information and evolutional patterns of a TKG simultaneously, considering the characteristics of real-world networks: community structure, scale-free and temporal decaying. First, we find communities in the input TKG to make the encoding get more similar intra-community embeddings. Second, we design a Hawkes process-based relational graph convolutional network to cope with the event impact-decaying phenomenon. Third, we design a conditional decoding method to alleviate biases towards frequent entities caused by long-tailed distribution. Experimental results show that HERLN achieves significant improvements over the state-of-the-art models.
2501.01980
Polarimetric BSSRDF Acquisition of Dynamic Faces
cs.CV cs.GR
Acquisition and modeling of polarized light reflection and scattering help reveal the shape, structure, and physical characteristics of an object, which is increasingly important in computer graphics. However, current polarimetric acquisition systems are limited to static and opaque objects. Human faces, on the other hand, present a particularly difficult challenge, given their complex structure and reflectance properties, the strong presence of spatially-varying subsurface scattering, and their dynamic nature. We present a new polarimetric acquisition method for dynamic human faces, which focuses on capturing spatially varying appearance and precise geometry, across a wide spectrum of skin tones and facial expressions. It includes both single and heterogeneous subsurface scattering, index of refraction, and specular roughness and intensity, among other parameters, while revealing biophysically-based components such as inner- and outer-layer hemoglobin, eumelanin and pheomelanin. Our method leverages such components' unique multispectral absorption profiles to quantify their concentrations, which in turn inform our model about the complex interactions occurring within the skin layers. To our knowledge, our work is the first to simultaneously acquire polarimetric and spectral reflectance information alongside biophysically-based skin parameters and geometry of dynamic human faces. Moreover, our polarimetric skin model integrates seamlessly into various rendering pipelines.
2501.01981
Optical Character Recognition using Convolutional Neural Networks for Ashokan Brahmi Inscriptions
cs.CV eess.IV
This research paper delves into the development of an Optical Character Recognition (OCR) system for the recognition of Ashokan Brahmi characters using Convolutional Neural Networks. It utilizes a comprehensive dataset of character images to train the models, along with data augmentation techniques to optimize the training process. Furthermore, the paper incorporates image preprocessing to remove noise, as well as image segmentation to facilitate line and character segmentation. The study mainly focuses on three pre-trained CNNs, namely LeNet, VGG-16, and MobileNet and compares their accuracy. Transfer learning was employed to adapt the pre-trained models to the Ashokan Brahmi character dataset. The findings reveal that MobileNet outperforms the other two models in terms of accuracy, achieving a validation accuracy of 95.94% and validation loss of 0.129. The paper provides an in-depth analysis of the implementation process using MobileNet and discusses the implications of the findings. The use of OCR for character recognition is of significant importance in the field of epigraphy, specifically for the preservation and digitization of ancient scripts. The results of this research paper demonstrate the effectiveness of using pre-trained CNNs for the recognition of Ashokan Brahmi characters.
2501.01982
Is Your Image a Good Storyteller?
cs.CV cs.AI cs.CL
Quantifying image complexity at the entity level is straightforward, but the assessment of semantic complexity has been largely overlooked. In fact, there are differences in semantic complexity across images. Images with richer semantics can tell vivid and engaging stories and offer a wide range of application scenarios. For example, the Cookie Theft picture is such a kind of image and is widely used to assess human language and cognitive abilities due to its higher semantic complexity. Additionally, semantically rich images can benefit the development of vision models, as images with limited semantics are becoming less challenging for them. However, such images are scarce, highlighting the need for a greater number of them. For instance, there is a need for more images like Cookie Theft to cater to people from different cultural backgrounds and eras. Assessing semantic complexity requires human experts and empirical evidence. Automatic evaluation of how semantically rich an image will be the first step of mining or generating more images with rich semantics, and benefit human cognitive assessment, Artificial Intelligence, and various other applications. In response, we propose the Image Semantic Assessment (ISA) task to address this problem. We introduce the first ISA dataset and a novel method that leverages language to solve this vision problem. Experiments on our dataset demonstrate the effectiveness of our approach. Our data and code are available at: https://github.com/xiujiesong/ISA.
2501.01983
ECG-guided individual identification via PPG
cs.CV cs.AI
Photoplethsmography (PPG)-based individual identification aiming at recognizing humans via intrinsic cardiovascular activities has raised extensive attention due to its high security and resistance to mimicry. However, this kind of technology witnesses unpromising results due to the limitation of low information density. To this end, electrocardiogram (ECG) signals have been introduced as a novel modality to enhance the density of input information. Specifically, a novel cross-modal knowledge distillation framework is implemented to propagate discriminate knowledge from ECG modality to PPG modality without incurring additional computational demands at the inference phase. Furthermore, to ensure efficient knowledge propagation, Contrastive Language-Image Pre-training (CLIP)-based knowledge alignment and cross-knowledge assessment modules are proposed respectively. Comprehensive experiments are conducted and results show our framework outperforms the baseline model with the improvement of 2.8% and 3.0% in terms of overall accuracy on seen- and unseen individual recognitions.
2501.01984
Leveraging AI for Automatic Classification of PCOS Using Ultrasound Imaging
eess.IV cs.AI cs.CV
The AUTO-PCOS Classification Challenge seeks to advance the diagnostic capabilities of artificial intelligence (AI) in identifying Polycystic Ovary Syndrome (PCOS) through automated classification of healthy and unhealthy ultrasound frames. This report outlines our methodology for building a robust AI pipeline utilizing transfer learning with the InceptionV3 architecture to achieve high accuracy in binary classification. Preprocessing steps ensured the dataset was optimized for training, validation, and testing, while interpretability methods like LIME and saliency maps provided valuable insights into the model's decision-making. Our approach achieved an accuracy of 90.52%, with precision, recall, and F1-score metrics exceeding 90% on validation data, demonstrating its efficacy. The project underscores the transformative potential of AI in healthcare, particularly in addressing diagnostic challenges like PCOS. Key findings, challenges, and recommendations for future enhancements are discussed, highlighting the pathway for creating reliable, interpretable, and scalable AI-driven medical diagnostic tools.
2501.01985
Fall Detection in Passenger Elevators using Intelligent Surveillance Camera Systems: An Application with YoloV8 Nano Model
cs.CV cs.AI
Computer vision technology, which involves analyzing images and videos captured by cameras through deep learning algorithms, has significantly advanced the field of human fall detection. This study focuses on the application of the YoloV8 Nano model in identifying fall incidents within passenger elevators, a context that presents unique challenges due to the enclosed environment and varying lighting conditions. By training the model on a robust dataset comprising over 10,000 images across diverse elevator types, we aim to enhance the detection precision and recall rates. The model's performance, with an 85% precision and 82% recall in fall detection, underscores its potential for integration into existing elevator safety systems to enable rapid intervention.
2501.01986
FrameFusion: Combining Similarity and Importance for Video Token Reduction on Large Visual Language Models
cs.CV cs.AI
The increasing demand to process long and high-resolution videos significantly burdens Large Vision-Language Models (LVLMs) due to the enormous number of visual tokens. Existing token reduction methods primarily focus on importance-based token pruning, which overlooks the redundancy caused by frame resemblance and repetitive visual elements. In this paper, we analyze the high vision token similarities in LVLMs. We reveal that token similarity distribution condenses as layers deepen while maintaining ranking consistency. Leveraging the unique properties of similarity over importance, we introduce FrameFusion, a novel approach that combines similarity-based merging with importance-based pruning for better token reduction in LVLMs. FrameFusion identifies and merges similar tokens before pruning, opening up a new perspective for token reduction. We evaluate FrameFusion on diverse LVLMs, including Llava-Video-{7B,32B,72B}, and MiniCPM-V-8B, on video understanding, question-answering, and retrieval benchmarks. Experiments show that FrameFusion reduces vision tokens by 70$\%$, achieving 3.4-4.4x LLM speedups and 1.6-1.9x end-to-end speedups, with an average performance impact of less than 3$\%$. Our code is available at https://github.com/thu-nics/FrameFusion.
2501.01987
Gender Bias in Text-to-Video Generation Models: A case study of Sora
cs.CV cs.AI cs.CY cs.LG
The advent of text-to-video generation models has revolutionized content creation as it produces high-quality videos from textual prompts. However, concerns regarding inherent biases in such models have prompted scrutiny, particularly regarding gender representation. Our study investigates the presence of gender bias in OpenAI's Sora, a state-of-the-art text-to-video generation model. We uncover significant evidence of bias by analyzing the generated videos from a diverse set of gender-neutral and stereotypical prompts. The results indicate that Sora disproportionately associates specific genders with stereotypical behaviors and professions, which reflects societal prejudices embedded in its training data.
2501.01989
CRRG-CLIP: Automatic Generation of Chest Radiology Reports and Classification of Chest Radiographs
cs.CV cs.AI
The complexity of stacked imaging and the massive number of radiographs make writing radiology reports complex and inefficient. Even highly experienced radiologists struggle to maintain accuracy and consistency in interpreting radiographs under prolonged high-intensity work. To address these issues, this work proposes the CRRG-CLIP Model (Chest Radiology Report Generation and Radiograph Classification Model), an end-to-end model for automated report generation and radiograph classification. The model consists of two modules: the radiology report generation module and the radiograph classification module. The generation module uses Faster R-CNN to identify anatomical regions in radiographs, a binary classifier to select key regions, and GPT-2 to generate semantically coherent reports. The classification module uses the unsupervised Contrastive Language Image Pretraining (CLIP) model, addressing the challenges of high-cost labelled datasets and insufficient features. The results show that the generation module performs comparably to high-performance baseline models on BLEU, METEOR, and ROUGE-L metrics, and outperformed the GPT-4o model on BLEU-2, BLEU-3, BLEU-4, and ROUGE-L metrics. The classification module significantly surpasses the state-of-the-art model in AUC and Accuracy. This demonstrates that the proposed model achieves high accuracy, readability, and fluency in report generation, while multimodal contrastive training with unlabelled radiograph-report pairs enhances classification performance.
2501.01990
Towards Sustainable Large Language Model Serving
cs.LG cs.DC
In this work, we study LLMs from a carbon emission perspective, addressing both operational and embodied emissions, and paving the way for sustainable LLM serving. We characterize the performance and energy of LLaMA with 1B, 3B, and 7B parameters using two Nvidia GPU types, a latest-generation RTX6000 Ada and an older-generation T4. We analytically model operational carbon emissions based on energy consumption and carbon intensities from three grid regions -- each representing a different energy source mix, and embodied carbon emissions based on chip area and memory size. Our characterization and modeling provide us with an in-depth understanding of the performance, energy, and carbon emissions of LLM serving. Our findings highlight the potential for optimizing sustainable LLM serving systems by considering both operational and embodied carbon emissions simultaneously.
2501.01991
A Hybrid Deep Learning and Model-Checking Framework for Accurate Brain Tumor Detection and Validation
cs.CV cs.AI
Model checking, a formal verification technique, ensures systems meet predefined requirements, playing a crucial role in minimizing errors and enhancing quality during development. This paper introduces a novel hybrid framework integrating model checking with deep learning for brain tumor detection and validation in medical imaging. By combining model-checking principles with CNN-based feature extraction and K-FCM clustering for segmentation, the proposed approach enhances the reliability of tumor detection and segmentation. Experimental results highlight the framework's effectiveness, achieving 98\% accuracy, 96.15\% precision, and 100\% recall, demonstrating its potential as a robust tool for advanced medical image analysis.
2501.01992
Disagree and Commit: Degrees of Argumentation-based Agreements
cs.AI cs.LO cs.MA
In cooperative human decision-making, agreements are often not total; a partial degree of agreement is sufficient to commit to a decision and move on, as long as one is somewhat confident that the involved parties are likely to stand by their commitment in the future, given no drastic unexpected changes. In this paper, we introduce the notion of agreement scenarios that allow artificial autonomous agents to reach such agreements, using formal models of argumentation, in particular abstract argumentation and value-based argumentation. We introduce the notions of degrees of satisfaction and (minimum, mean, and median) agreement, as well as a measure of the impact a value in a value-based argumentation framework has on these notions. We then analyze how degrees of agreement are affected when agreement scenarios are expanded with new information, to shed light on the reliability of partial agreements in dynamic scenarios. An implementation of the introduced concepts is provided as part of an argumentation-based reasoning software library.
2501.01993
A Novel Convolution and Attention Mechanism-based Model for 6D Object Pose Estimation
cs.CV cs.LG
Estimating 6D object poses from RGB images is challenging because the lack of depth information requires inferring a three dimensional structure from 2D projections. Traditional methods often rely on deep learning with grid based data structures but struggle to capture complex dependencies among extracted features. To overcome this, we introduce a graph based representation derived directly from images, where spatial temporal features of each pixel serve as nodes, and relationships between them are defined through node connectivity and spatial interactions. We also employ feature selection mechanisms that use spatial attention and self attention distillation, along with a Legendre convolution layer leveraging the orthogonality of Legendre polynomials for numerical stability. Experiments on the LINEMOD, Occluded LINEMOD, and YCB Video datasets demonstrate that our method outperforms nine existing approaches and achieves state of the art benchmark in object pose estimation.
2501.01994
Fuzzy Model Identification and Self Learning with Smooth Compositions
eess.SY cs.AI cs.SY
This paper develops a smooth model identification and self-learning strategy for dynamic systems taking into account possible parameter variations and uncertainties. We have tried to solve the problem such that the model follows the changes and variations in the system on a continuous and smooth surface. Running the model to adaptively gain the optimum values of the parameters on a smooth surface would facilitate further improvements in the application of other derivative based optimization control algorithms such as MPC or robust control algorithms to achieve a combined modeling-control scheme. Compared to the earlier works on the smooth fuzzy modeling structures, we could reach a desired trade-off between the model optimality and the computational load. The proposed method has been evaluated on a test problem as well as the non-linear dynamic of a chemical process.
2501.01998
SmartSpatial: Enhancing the 3D Spatial Arrangement Capabilities of Stable Diffusion Models and Introducing a Novel 3D Spatial Evaluation Framework
cs.CV cs.AI
Stable Diffusion models have made remarkable strides in generating photorealistic images from text prompts but often falter when tasked with accurately representing complex spatial arrangements, particularly involving intricate 3D relationships. To address this limitation, we introduce SmartSpatial, an innovative approach that enhances the spatial arrangement capabilities of Stable Diffusion models through 3D-aware conditioning and attention-guided mechanisms. SmartSpatial incorporates depth information and employs cross-attention control to ensure precise object placement, delivering notable improvements in spatial accuracy metrics. In conjunction with SmartSpatial, we present SmartSpatialEval, a comprehensive evaluation framework designed to assess spatial relationships. This framework utilizes vision-language models and graph-based dependency parsing for performance analysis. Experimental results on the COCO and SpatialPrompts datasets show that SmartSpatial significantly outperforms existing methods, setting new benchmarks for spatial arrangement accuracy in image generation.
2501.01999
On the Utility of Equivariance and Symmetry Breaking in Deep Learning Architectures on Point Clouds
cs.CV cs.AI cs.LG
This paper explores the key factors that influence the performance of models working with point clouds, across different tasks of varying geometric complexity. In this work, we explore the trade-offs between flexibility and weight-sharing introduced by equivariant layers, assessing when equivariance boosts or detracts from performance. It is often argued that providing more information as input improves a model's performance. However, if this additional information breaks certain properties, such as $\SE(3)$ equivariance, does it remain beneficial? We identify the key aspects of equivariant and non-equivariant architectures that drive success in different tasks by benchmarking them on segmentation, regression, and generation tasks across multiple datasets with increasing complexity. We observe a positive impact of equivariance, which becomes more pronounced with increasing task complexity, even when strict equivariance is not required.
2501.02000
Multi-Center Study on Deep Learning-Assisted Detection and Classification of Fetal Central Nervous System Anomalies Using Ultrasound Imaging
eess.IV cs.AI cs.CV
Prenatal ultrasound evaluates fetal growth and detects congenital abnormalities during pregnancy, but the examination of ultrasound images by radiologists requires expertise and sophisticated equipment, which would otherwise fail to improve the rate of identifying specific types of fetal central nervous system (CNS) abnormalities and result in unnecessary patient examinations. We construct a deep learning model to improve the overall accuracy of the diagnosis of fetal cranial anomalies to aid prenatal diagnosis. In our collected multi-center dataset of fetal craniocerebral anomalies covering four typical anomalies of the fetal central nervous system (CNS): anencephaly, encephalocele (including meningocele), holoprosencephaly, and rachischisis, patient-level prediction accuracy reaches 94.5%, with an AUROC value of 99.3%. In the subgroup analyzes, our model is applicable to the entire gestational period, with good identification of fetal anomaly types for any gestational period. Heatmaps superimposed on the ultrasound images not only provide a visual interpretation for the algorithm but also provide an intuitive visual aid to the physician by highlighting key areas that need to be reviewed, helping the physician to quickly identify and validate key areas. Finally, the retrospective reader study demonstrates that by combining the automatic prediction of the DL system with the professional judgment of the radiologist, the diagnostic accuracy and efficiency can be effectively improved and the misdiagnosis rate can be reduced, which has an important clinical application prospect.
2501.02001
Communication Efficient Cooperative Edge AI via Event-Triggered Computation Offloading
cs.LG eess.IV eess.SP
Rare events, despite their infrequency, often carry critical information and require immediate attentions in mission-critical applications such as autonomous driving, healthcare, and industrial automation. The data-intensive nature of these tasks and their need for prompt responses, combined with designing edge AI (or edge inference), pose significant challenges in systems and techniques. Existing edge inference approaches often suffer from communication bottlenecks due to high-dimensional data transmission and fail to provide timely responses to rare events, limiting their effectiveness for mission-critical applications in the sixth-generation (6G) mobile networks. To overcome these challenges, we propose a channel-adaptive, event-triggered edge-inference framework that prioritizes efficient rare-event processing. Central to this framework is a dual-threshold, multi-exit architecture, which enables early local inference for rare events detected locally while offloading more complex rare events to edge servers for detailed classification. To further enhance the system's performance, we developed a channel-adaptive offloading policy paired with an online algorithm to dynamically determine the optimal confidence thresholds for controlling offloading decisions. The associated optimization problem is solved by reformulating the original non-convex function into an equivalent strongly convex one. Using deep neural network classifiers and real medical datasets, our experiments demonstrate that the proposed framework not only achieves superior rare-event classification accuracy, but also effectively reduces communication overhead, as opposed to existing edge-inference approaches.