id
stringlengths
9
16
title
stringlengths
4
278
categories
stringlengths
5
104
abstract
stringlengths
6
4.09k
2501.16312
LinPrim: Linear Primitives for Differentiable Volumetric Rendering
cs.CV
Volumetric rendering has become central to modern novel view synthesis methods, which use differentiable rendering to optimize 3D scene representations directly from observed views. While many recent works build on NeRF or 3D Gaussians, we explore an alternative volumetric scene representation. More specifically, we introduce two new scene representations based on linear primitives-octahedra and tetrahedra-both of which define homogeneous volumes bounded by triangular faces. This formulation aligns naturally with standard mesh-based tools, minimizing overhead for downstream applications. To optimize these primitives, we present a differentiable rasterizer that runs efficiently on GPUs, allowing end-to-end gradient-based optimization while maintaining realtime rendering capabilities. Through experiments on real-world datasets, we demonstrate comparable performance to state-of-the-art volumetric methods while requiring fewer primitives to achieve similar reconstruction fidelity. Our findings provide insights into the geometry of volumetric rendering and suggest that adopting explicit polyhedra can expand the design space of scene representations.
2501.16319
Adaptive Iterative Compression for High-Resolution Files: an Approach Focused on Preserving Visual Quality in Cinematic Workflows
cs.CV cs.ET cs.LG cs.PF
This study presents an iterative adaptive compression model for high-resolution DPX-derived TIFF files used in cinematographic workflows and digital preservation. The model employs SSIM and PSNR metrics to dynamically adjust compression parameters across three configurations (C0, C1, C2), achieving storage reductions up to 83.4 % while maintaining high visual fidelity (SSIM > 0.95). Validation across three diverse productions - black and white classic, soft-palette drama, and complex action film - demonstrated the method's effectiveness in preserving critical visual elements while significantly reducing storage requirements. Professional evaluators reported 90% acceptance rate for the optimal C1 configuration, with artifacts remaining below perceptual threshold in critical areas. Comparative analysis with JPEG2000 and H.265 showed superior quality preservation at equivalent compression rates, particularly for high bit-depth content. While requiring additional computational overhead, the method's storage benefits and quality control capabilities make it suitable for professional workflows, with potential applications in medical imaging and cloud storage optimization.
2501.16322
Implicit Bias in Matrix Factorization and its Explicit Realization in a New Architecture
cs.LG math.OC stat.ML
Gradient descent for matrix factorization is known to exhibit an implicit bias toward approximately low-rank solutions. While existing theories often assume the boundedness of iterates, empirically the bias persists even with unbounded sequences. We thus hypothesize that implicit bias is driven by divergent dynamics markedly different from the convergent dynamics for data fitting. Using this perspective, we introduce a new factorization model: $X\approx UDV^\top$, where $U$ and $V$ are constrained within norm balls, while $D$ is a diagonal factor allowing the model to span the entire search space. Our experiments reveal that this model exhibits a strong implicit bias regardless of initialization and step size, yielding truly (rather than approximately) low-rank solutions. Furthermore, drawing parallels between matrix factorization and neural networks, we propose a novel neural network model featuring constrained layers and diagonal components. This model achieves strong performance across various regression and classification tasks while finding low-rank solutions, resulting in efficient and lightweight networks.
2501.16325
Tailored Forecasting from Short Time Series via Meta-learning
cs.LG nlin.CD physics.comp-ph
Machine learning (ML) models can be effective for forecasting the dynamics of unknown systems from time-series data, but they often require large amounts of data and struggle to generalize across systems with varying dynamics. Combined, these issues make forecasting from short time series particularly challenging. To address this problem, we introduce Meta-learning for Tailored Forecasting from Related Time Series (METAFORS), which uses related systems with longer time-series data to supplement limited data from the system of interest. By leveraging a library of models trained on related systems, METAFORS builds tailored models to forecast system evolution with limited data. Using a reservoir computing implementation and testing on simulated chaotic systems, we demonstrate METAFORS' ability to predict both short-term dynamics and long-term statistics, even when test and related systems exhibit significantly different behaviors and the available data are scarce, highlighting its robustness and versatility in data-limited scenarios.
2501.16327
LUCY: Linguistic Understanding and Control Yielding Early Stage of Her
cs.CL cs.SD eess.AS
The film Her features Samantha, a sophisticated AI audio agent who is capable of understanding both linguistic and paralinguistic information in human speech and delivering real-time responses that are natural, informative and sensitive to emotional subtleties. Moving one step toward more sophisticated audio agent from recent advancement in end-to-end (E2E) speech systems, we propose LUCY, a E2E speech model that (1) senses and responds to user's emotion, (2) deliver responses in a succinct and natural style, and (3) use external tool to answer real-time inquiries. Experiment results show that LUCY is better at emotion control than peer models, generating emotional responses based on linguistic emotional instructions and responding to paralinguistic emotional cues. Lucy is also able to generate responses in a more natural style, as judged by external language models, without sacrificing much performance on general question answering. Finally, LUCY can leverage function calls to answer questions that are out of its knowledge scope.
2501.16329
sDREAMER: Self-distilled Mixture-of-Modality-Experts Transformer for Automatic Sleep Staging
cs.LG cs.AI
Automatic sleep staging based on electroencephalography (EEG) and electromyography (EMG) signals is an important aspect of sleep-related research. Current sleep staging methods suffer from two major drawbacks. First, there are limited information interactions between modalities in the existing methods. Second, current methods do not develop unified models that can handle different sources of input. To address these issues, we propose a novel sleep stage scoring model sDREAMER, which emphasizes cross-modality interaction and per-channel performance. Specifically, we develop a mixture-of-modality-expert (MoME) model with three pathways for EEG, EMG, and mixed signals with partially shared weights. We further propose a self-distillation training scheme for further information interaction across modalities. Our model is trained with multi-channel inputs and can make classifications on either single-channel or multi-channel inputs. Experiments demonstrate that our model outperforms the existing transformer-based sleep scoring methods for multi-channel inference. For single-channel inference, our model also outperforms the transformer-based models trained with single-channel signals.
2501.16330
RelightVid: Temporal-Consistent Diffusion Model for Video Relighting
cs.CV cs.AI
Diffusion models have demonstrated remarkable success in image generation and editing, with recent advancements enabling albedo-preserving image relighting. However, applying these models to video relighting remains challenging due to the lack of paired video relighting datasets and the high demands for output fidelity and temporal consistency, further complicated by the inherent randomness of diffusion models. To address these challenges, we introduce RelightVid, a flexible framework for video relighting that can accept background video, text prompts, or environment maps as relighting conditions. Trained on in-the-wild videos with carefully designed illumination augmentations and rendered videos under extreme dynamic lighting, RelightVid achieves arbitrary video relighting with high temporal consistency without intrinsic decomposition while preserving the illumination priors of its image backbone.
2501.16331
Decoding OTC Government Bond Market Liquidity: An ABM Model for Market Dynamics
q-fin.TR cs.AI
The over-the-counter (OTC) government bond markets are characterised by their bilateral trading structures, which pose unique challenges to understanding and ensuring market stability and liquidity. In this paper, we develop a bespoke ABM that simulates market-maker interactions within a stylised government bond market. The model focuses on the dynamics of liquidity and stability in the secondary trading of government bonds, particularly in concentrated markets like those found in Australia and the UK. Through this simulation, we test key hypotheses around improving market stability, focusing on the effects of agent diversity, business costs, and client base size. We demonstrate that greater agent diversity enhances market liquidity and that reducing the costs of market-making can improve overall market stability. The model offers insights into computational finance by simulating trading without price transparency, highlighting how micro-structural elements can affect macro-level market outcomes. This research contributes to the evolving field of computational finance by employing computational intelligence techniques to better understand the fundamental mechanics of government bond markets, providing actionable insights for both academics and practitioners.
2501.16333
A New Proof for the Linear Filtering and Smoothing Equations, and Asymptotic Expansion of Nonlinear Filtering
eess.SP cs.IT math.IT math.PR math.ST stat.TH
In this paper, we propose a new approach to the linear filtering and smoothing problem and demonstrate its applicability to nonlinear filtering. For the linear case, our main theorem provides an explicit expression for the conditional distribution of the hidden process given the observations, leading to a novel derivation of the linear filtering and smoothing equations. Moreover, the theorem offers an efficient framework for computing the asymptotic expansion of nonlinear filtering.
2501.16334
RNN-Based Models for Predicting Seizure Onset in Epileptic Patients
eess.SP cs.LG
Early management and better clinical outcomes for epileptic patients depend on seizure prediction. The accuracy and false alarm rates of existing systems are often compromised by their dependence on static thresholds and basic Electroencephalogram (EEG) properties. A novel Recurrent Neural Network (RNN)-based method for seizure start prediction is proposed in the article to overcome these limitations. As opposed to conventional techniques, the proposed system makes use of Long Short-Term Memory (LSTM) networks to extract temporal correlations from unprocessed EEG data. It enables the system to adapt dynamically to the unique EEG patterns of each patient, improving prediction accuracy. The methodology of the system comprises thorough data collecting, preprocessing, and LSTM-based feature extraction. Annotated EEG datasets are then used for model training and validation. Results show a considerable reduction in false alarm rates (average of 6.8%) and an improvement in prediction accuracy (90.2% sensitivity, 88.9% specificity, and AUC-ROC of 93). Additionally, computational efficiency is significantly higher than that of existing systems (12 ms processing time, 45 MB memory consumption). About improving seizure prediction reliability, these results demonstrate the effectiveness of the proposed RNN-based strategy, opening up possibilities for its practical application to improve epilepsy treatment.
2501.16336
Runtime Analysis of Evolutionary Algorithms for Multiparty Multiobjective Optimization
cs.NE cs.AI
In scenarios where multiple decision-makers operate within a common decision space, each focusing on their own multi-objective optimization problem (e.g., bargaining games), the problem can be modeled as a multi-party multi-objective optimization problem (MPMOP). While numerous evolutionary algorithms have been proposed to solve MPMOPs, most results remain empirical. This paper presents the first theoretical analysis of the expected runtime of evolutionary algorithms on bi-party multi-objective optimization problems (BPMOPs). Our findings demonstrate that employing traditional multi-objective optimization algorithms to solve MPMOPs is both time-consuming and inefficient, as the resulting population contains many solutions that fail to achieve consensus among decision-makers. An alternative approach involves decision-makers individually solving their respective optimization problems and seeking consensus only in the final stage. While feasible for pseudo-Boolean optimization problems, this method may fail to guarantee approximate performance for one party in NP-hard problems. Finally, We propose coevolutionary multi-party multi-objective optimizers (CoEMPMO) for pseudo-Boolean optimization and shortest path problems within a multi-party multi-objective context, which maintains a common solution set among all parties through coevolution. Theoretical and experimental results demonstrate that the proposed \( \text{CoEMPMO}_{\text{random}} \) outperforms previous algorithms in terms of the expected lower bound on runtime for pseudo-Boolean optimization problems. Additionally, \( \text{CoEMPMO}_{\text{cons}}^{\text{SP}} \) achieves better efficiency and precision in solving shortest path problems compared to existing algorithms.
2501.16337
Explore Activation Sparsity in Recurrent LLMs for Energy-Efficient Neuromorphic Computing
cs.NE cs.AI cs.AR cs.LG
The recent rise of Large Language Models (LLMs) has revolutionized the deep learning field. However, the desire to deploy LLMs on edge devices introduces energy efficiency and latency challenges. Recurrent LLM (R-LLM) architectures have proven effective in mitigating the quadratic complexity of self-attention, making them a potential paradigm for computing on-edge neuromorphic processors. In this work, we propose a low-cost, training-free algorithm to sparsify R-LLMs' activations to enhance energy efficiency on neuromorphic hardware. Our approach capitalizes on the inherent structure of these models, rendering them well-suited for energy-constrained environments. Although primarily designed for R-LLMs, this method can be generalized to other LLM architectures, such as transformers, as demonstrated on the OPT model, achieving comparable sparsity and efficiency improvements. Empirical studies illustrate that our method significantly reduces computational demands while maintaining competitive accuracy across multiple zero-shot learning benchmarks. Additionally, hardware simulations with the SENECA neuromorphic processor underscore notable energy savings and latency improvements. These results pave the way for low-power, real-time neuromorphic deployment of LLMs and demonstrate the feasibility of training-free on-chip adaptation using activation sparsity.
2501.16341
Developing Enhanced Conversational Agents for Social Virtual Worlds
eess.AS cs.CL cs.SD
In this paper, we present a methodology for the development of embodied conversational agents for social virtual worlds. The agents provide multimodal communication with their users in which speech interaction is included. Our proposal combines different techniques related to Artificial Intelligence, Natural Language Processing, Affective Computing, and User Modeling. Firstly, the developed conversational agents. A statistical methodology has been developed to model the system conversational behavior, which is learned from an initial corpus and improved with the knowledge acquired from the successive interactions. In addition, the selection of the next system response is adapted considering information stored into users profiles and also the emotional contents detected in the users utterances. Our proposal has been evaluated with the successful development of an embodied conversational agent which has been placed in the Second Life social virtual world. The avatar includes the different models and interacts with the users who inhabit the virtual world in order to provide academic information. The experimental results show that the agents conversational behavior adapts successfully to the specific characteristics of users interacting in such environments.
2501.16343
Self-orthogonal and self-dual codes from maximal curves
cs.IT math.AG math.IT
In the field of algebraic geometric codes (AG codes), the characterization of dual codes has long been a challenging problem which relies on differentials. In this paper, we provide some descriptions for certain differentials utilizing algebraic structure of finite fields and geometric properties of algebraic curves. Moreover, we construct self-orthogonal and self-dual codes with parameters $[n, k, d]_{q^2}$ satisfying $k + d$ is close to $n$. Additionally, quantum codes with large minimum distance are also constructed.
2501.16344
WhiSPA: Semantically and Psychologically Aligned Whisper with Self-Supervised Contrastive and Student-Teacher Learning
eess.AS cs.AI cs.CL cs.SD
Current speech encoding pipelines often rely on an additional text-based LM to get robust representations of human communication, even though SotA speech-to-text models often have a LM within. This work proposes an approach to improve the LM within an audio model such that the subsequent text-LM is unnecessary. We introduce WhiSPA (Whisper with Semantic and Psychological Alignment), which leverages a novel audio training objective: contrastive loss with a language model embedding as a teacher. Using over 500k speech segments from mental health audio interviews, we evaluate the utility of aligning Whisper's latent space with semantic representations from a text autoencoder (SBERT) and lexically derived embeddings of basic psychological dimensions: emotion and personality. Over self-supervised affective tasks and downstream psychological tasks, WhiSPA surpasses current speech encoders, achieving an average error reduction of 73.4% and 83.8%, respectively. WhiSPA demonstrates that it is not always necessary to run a subsequent text LM on speech-to-text output in order to get a rich psychological representation of human communication.
2501.16345
Self-Clustering Graph Transformer Approach to Model Resting-State Functional Brain Activity
cs.LG cs.AI
Resting-state functional magnetic resonance imaging (rs-fMRI) offers valuable insights into the human brain's functional organization and is a powerful tool for investigating the relationship between brain function and cognitive processes, as it allows for the functional organization of the brain to be captured without relying on a specific task or stimuli. In this study, we introduce a novel attention mechanism for graphs with subnetworks, named Self-Clustering Graph Transformer (SCGT), designed to handle the issue of uniform node updates in graph transformers. By using static functional connectivity (FC) correlation features as input to the transformer model, SCGT effectively captures the sub-network structure of the brain by performing cluster-specific updates to the nodes, unlike uniform node updates in vanilla graph transformers, further allowing us to learn and interpret the subclusters. We validate our approach on the Adolescent Brain Cognitive Development (ABCD) dataset, comprising 7,957 participants, for the prediction of total cognitive score and gender classification. Our results demonstrate that SCGT outperforms the vanilla graph transformer method and other recent models, offering a promising tool for modeling brain functional connectivity and interpreting the underlying subnetwork structures.
2501.16346
Self-supervised Graph Transformer with Contrastive Learning for Brain Connectivity Analysis towards Improving Autism Detection
cs.LG cs.AI
Functional Magnetic Resonance Imaging (fMRI) provides useful insights into the brain function both during task or rest. Representing fMRI data using correlation matrices is found to be a reliable method of analyzing the inherent connectivity of the brain in the resting and active states. Graph Neural Networks (GNNs) have been widely used for brain network analysis due to their inherent explainability capability. In this work, we introduce a novel framework using contrastive self-supervised learning graph transformers, incorporating a brain network transformer encoder with random graph alterations. The proposed network leverages both contrastive learning and graph alterations to effectively train the graph transformer for autism detection. Our approach, tested on Autism Brain Imaging Data Exchange (ABIDE) data, demonstrates superior autism detection, achieving an AUROC of 82.6 and an accuracy of 74%, surpassing current state-of-the-art methods.
2501.16347
Identification of Hardware Trojan Locations in Gate-Level Netlist using Nearest Neighbour Approach integrated with Machine Learning Technique
cs.LG cs.AI
In the evolving landscape of integrated circuit design, detecting Hardware Trojans (HTs) within a multi entity based design cycle presents significant challenges. This research proposes an innovative machine learning-based methodology for identifying malicious logic gates in gate-level netlists. By focusing on path retrace algorithms. The methodology is validated across three distinct cases, each employing different machine learning models to classify HTs. Case I utilizes a decision tree algorithm for node-to-node comparisons, significantly improving detection accuracy through the integration of Principal Component Analysis (PCA). Case II introduces a graph-to-graph classification using a Graph Neural Network (GNN) model, enabling the differentiation between normal and Trojan-infected circuit designs. Case III applies GNN-based node classification to identify individual compromised nodes and its location. Additionally, nearest neighbor (NN) method has been combined with GNN graph-to-graph in Case II and GNN node-to-node in Case III. Despite the potential of GNN model graph-to-graph classification, NN approach demonstrated superior performance, with the first nearest neighbor (1st NN) achieving 73.2% accuracy and the second nearest neighbor (2nd NN) method reaching 97.7%. In comparison, the GNN model achieved an accuracy of 62.8%. Similarly, GNN model node-to-node classification, NN approach demonstrated superior performance, with the 1st NN achieving 93% accuracy and the 2nd NN method reaching 97.7%. In comparison, the GNN model achieved an accuracy of 79.8%. However, higher and higher NN will lead to large code coverage for the identification of HTs.
2501.16348
An Integrated Approach to AI-Generated Content in e-health
cs.LG cs.AI
Artificial Intelligence-Generated Content, a subset of Generative Artificial Intelligence, holds significant potential for advancing the e-health sector by generating diverse forms of data. In this paper, we propose an end-to-end class-conditioned framework that addresses the challenge of data scarcity in health applications by generating synthetic medical images and text data, evaluating on practical applications such as retinopathy detection, skin infections and mental health assessments. Our framework integrates Diffusion and Large Language Models (LLMs) to generate data that closely match real-world patterns, which is essential for improving downstream task performance and model robustness in e-health applications. Experimental results demonstrate that the synthetic images produced by the proposed diffusion model outperform traditional GAN architectures. Similarly, in the text modality, data generated by uncensored LLM achieves significantly better alignment with real-world data than censored models in replicating the authentic tone.
2501.16349
Risk-Informed Diffusion Transformer for Long-Tail Trajectory Prediction in the Crash Scenario
cs.LG cs.AI
Trajectory prediction methods have been widely applied in autonomous driving technologies. Although the overall performance accuracy of trajectory prediction is relatively high, the lack of trajectory data in critical scenarios in the training data leads to the long-tail phenomenon. Normally, the trajectories of the tail data are more critical and more difficult to predict and may include rare scenarios such as crashes. To solve this problem, we extracted the trajectory data from real-world crash scenarios, which contain more long-tail data. Meanwhile, based on the trajectory data in this scenario, we integrated graph-based risk information and diffusion with transformer and proposed the Risk-Informed Diffusion Transformer (RI-DiT) trajectory prediction method. Extensive experiments were conducted on trajectory data in the real-world crash scenario, and the results show that the algorithm we proposed has good performance. When predicting the data of the tail 10\% (Top 10\%), the minADE and minFDE indicators are 0.016/2.667 m. At the same time, we showed the trajectory conditions of different long-tail distributions. The distribution of trajectory data is closer to the tail, the less smooth the trajectory is. Through the trajectory data in real-world crash scenarios, Our work expands the methods to overcome the long-tail challenges in trajectory prediction. Our method, RI-DiT, integrates inverse time to collision (ITTC) and the feature of traffic flow, which can predict long-tail trajectories more accurately and improve the safety of autonomous driving systems.
2501.16350
A Method for Multi-Hop Question Answering on Persian Knowledge Graph
cs.IR cs.AI cs.CL
Question answering systems are the latest evolution in information retrieval technology, designed to accept complex queries in natural language and provide accurate answers using both unstructured and structured knowledge sources. Knowledge Graph Question Answering (KGQA) systems fulfill users' information needs by utilizing structured data, representing a vast number of facts as a graph. However, despite significant advancements, major challenges persist in answering multi-hop complex questions, particularly in Persian. One of the main challenges is the accurate understanding and transformation of these multi-hop complex questions into semantically equivalent SPARQL queries, which allows for precise answer retrieval from knowledge graphs. In this study, to address this issue, a dataset of 5,600 Persian multi-hop complex questions was developed, along with their decomposed forms based on the semantic representation of the questions. Following this, Persian language models were trained using this dataset, and an architecture was proposed for answering complex questions using a Persian knowledge graph. Finally, the proposed method was evaluated against similar systems on the PeCoQ dataset. The results demonstrated the superiority of our approach, with an improvement of 12.57% in F1-score and 12.06% in accuracy compared to the best comparable method.
2501.16352
Mixture of Experts (MoE): A Big Data Perspective
cs.LG cs.AI
As the era of big data arrives, traditional artificial intelligence algorithms have difficulty processing the demands of massive and diverse data. Mixture of experts (MoE) has shown excellent performance and broad application prospects. This paper provides an in-depth review and analysis of the latest progress in this field from multiple perspectives, including the basic principles, algorithmic models, key technical challenges, and application practices of MoE. First, we introduce the basic concept of MoE and its core idea and elaborate on its advantages over traditional single models. Then, we discuss the basic architecture of MoE and its main components, including the gating network, expert networks, and learning algorithms. Next, we review the applications of MoE in addressing key technical issues in big data. For each challenge, we provide specific MoE solutions and their innovations. Furthermore, we summarize the typical use cases of MoE in various application domains. This fully demonstrates the powerful capability of MoE in big data processing. We also analyze the advantages of MoE in big data environments. Finally, we explore the future development trends of MoE. We believe that MoE will become an important paradigm of artificial intelligence in the era of big data. In summary, this paper systematically elaborates on the principles, techniques, and applications of MoE in big data processing, providing theoretical and practical references to further promote the application of MoE in real scenarios.
2501.16353
Synthetic Data Generation by Supervised Neural Gas Network for Physiological Emotion Recognition Data
cs.NE cs.AI cs.LG eess.SP
Data scarcity remains a significant challenge in the field of emotion recognition using physiological signals, as acquiring comprehensive and diverse datasets is often prevented by privacy concerns and logistical constraints. This limitation restricts the development and generalization of robust emotion recognition models, making the need for effective synthetic data generation methods more critical. Emotion recognition from physiological signals such as EEG, ECG, and GSR plays a pivotal role in enhancing human-computer interaction and understanding human affective states. Utilizing these signals, this study introduces an innovative approach to synthetic data generation using a Supervised Neural Gas (SNG) network, which has demonstrated noteworthy speed advantages over established models like Conditional VAE, Conditional GAN, diffusion model, and Variational LSTM. The Neural Gas network, known for its adaptability in organizing data based on topological and feature-space proximity, provides a robust framework for generating real-world-like synthetic datasets that preserve the intrinsic patterns of physiological emotion data. Our implementation of the SNG efficiently processes the input data, creating synthetic instances that closely mimic the original data distributions, as demonstrated through comparative accuracy assessments. In experiments, while our approach did not universally outperform all models, it achieved superior performance against most of the evaluated models and offered significant improvements in processing time. These outcomes underscore the potential of using SNG networks for fast, efficient, and effective synthetic data generation in emotion recognition applications.
2501.16354
Adaptive Hoeffding Tree with Transfer Learning for Streaming Synchrophasor Data Sets
cs.LG cs.AI
Synchrophasor technology or phasor measurement units (PMUs) are known to detect multiple type of oscillations or faults better than Supervisory Control and Data Acquisition (SCADA) systems, but the volume of Bigdata (e.g., 30-120 samples per second on a single PMU) generated by these sensors at the aggregator level (e.g., several PMUs) requires special handling. Conventional machine learning or data mining methods are not suitable to handle such larger streaming realtime data. This is primarily due to latencies associated with cloud environments (e.g., at an aggregator or PDC level), and thus necessitates the need for local computing to move the data on the edge (or locally at the PMU level) for processing. This requires faster real-time streaming algorithms to be processed at the local level (e.g., typically by a Field Programmable Gate Array (FPGA) based controllers). This paper proposes a transfer learning-based hoeffding tree with ADWIN (THAT) method to detect anomalous synchrophasor signatures. The proposed algorithm is trained and tested with the OzaBag method. The preliminary results with transfer learning indicate that a computational time saving of 0.7ms is achieved with THAT algorithm (0.34ms) over Ozabag (1.04ms), while the accuracy of both methods in detecting fault events remains at 94% for four signatures.
2501.16355
How Strategic Agents Respond: Comparing Analytical Models with LLM-Generated Responses in Strategic Classification
cs.LG cs.AI
When machine learning (ML) algorithms are used to automate human-related decisions, human agents may gain knowledge of the decision policy and behave strategically to obtain desirable outcomes. Strategic Classification (SC) has been proposed to address the interplay between agents and decision-makers. Prior work on SC has relied on assumptions that agents are perfectly or approximately rational, responding to decision policies by maximizing their utilities. Verifying these assumptions is challenging due to the difficulty of collecting real-world agent responses. Meanwhile, the growing adoption of large language models (LLMs) makes it increasingly likely that human agents in SC settings will seek advice from these tools. We propose using strategic advice generated by LLMs to simulate human agent responses in SC. Specifically, we examine five critical SC scenarios -- hiring, loan applications, school admissions, personal income, and public assistance programs -- and simulate how human agents with diverse profiles seek advice from LLMs. We then compare the resulting agent responses with the best responses generated by existing theoretical models. Our findings reveal that: (i) LLMs and theoretical models generally lead to agent score or qualification changes in the same direction across most settings, with both achieving similar levels of fairness; (ii) state-of-the-art commercial LLMs (e.g., GPT-3.5, GPT-4) consistently provide helpful suggestions, though these suggestions typically do not result in maximal score or qualification improvements; and (iii) LLMs tend to produce more diverse agent responses, often favoring more balanced effort allocation strategies. These results suggest that theoretical models align with LLMs to some extent and that leveraging LLMs to simulate more realistic agent responses offers a promising approach to designing trustworthy ML systems.
2501.16356
Evaluating Binary Decision Biases in Large Language Models: Implications for Fair Agent-Based Financial Simulations
cs.LG cs.AI
Large Language Models (LLMs) are increasingly being used to simulate human-like decision making in agent-based financial market models (ABMs). As models become more powerful and accessible, researchers can now incorporate individual LLM decisions into ABM environments. However, integration may introduce inherent biases that need careful evaluation. In this paper we test three state-of-the-art GPT models for bias using two model sampling approaches: one-shot and few-shot API queries. We observe significant variations in distributions of outputs between specific models, and model sub versions, with GPT-4o-Mini-2024-07-18 showing notably better performance (32-43% yes responses) compared to GPT-4-0125-preview's extreme bias (98-99% yes responses). We show that sampling methods and model sub-versions significantly impact results: repeated independent API calls produce different distributions compared to batch sampling within a single call. While no current GPT model can simultaneously achieve a uniform distribution and Markovian properties in one-shot testing, few-shot sampling can approach uniform distributions under certain conditions. We explore the Temperature parameter, providing a definition and comparative results. We further compare our results to true random binary series and test specifically for the common human bias of Negative Recency - finding LLMs have a mixed ability to 'beat' humans in this one regard. These findings emphasise the critical importance of careful LLM integration into ABMs for financial markets and more broadly.
2501.16357
EVolutionary Independent DEtermiNistiC Explanation
cs.LG cs.AI eess.SP
The widespread use of artificial intelligence deep neural networks in fields such as medicine and engineering necessitates understanding their decision-making processes. Current explainability methods often produce inconsistent results and struggle to highlight essential signals influencing model inferences. This paper introduces the Evolutionary Independent Deterministic Explanation (EVIDENCE) theory, a novel approach offering a deterministic, model-independent method for extracting significant signals from black-box models. EVIDENCE theory, grounded in robust mathematical formalization, is validated through empirical tests on diverse datasets, including COVID-19 audio diagnostics, Parkinson's disease voice recordings, and the George Tzanetakis music classification dataset (GTZAN). Practical applications of EVIDENCE include improving diagnostic accuracy in healthcare and enhancing audio signal analysis. For instance, in the COVID-19 use case, EVIDENCE-filtered spectrograms fed into a frozen Residual Network with 50 layers improved precision by 32% for positive cases and increased the area under the curve (AUC) by 16% compared to baseline models. For Parkinson's disease classification, EVIDENCE achieved near-perfect precision and sensitivity, with a macro average F1-Score of 0.997. In the GTZAN, EVIDENCE maintained a high AUC of 0.996, demonstrating its efficacy in filtering relevant features for accurate genre classification. EVIDENCE outperformed other Explainable Artificial Intelligence (XAI) methods such as LIME, SHAP, and GradCAM in almost all metrics. These findings indicate that EVIDENCE not only improves classification accuracy but also provides a transparent and reproducible explanation mechanism, crucial for advancing the trustworthiness and applicability of AI systems in real-world settings.
2501.16358
The OpenLAM Challenges
cs.LG cond-mat.mtrl-sci physics.comp-ph
Inspired by the success of Large Language Models (LLMs), the development of Large Atom Models (LAMs) has gained significant momentum in scientific computation. Since 2022, the Deep Potential team has been actively pretraining LAMs and launched the OpenLAM Initiative to develop an open-source foundation model spanning the periodic table. A core objective is establishing comprehensive benchmarks for reliable LAM evaluation, addressing limitations in existing datasets. As a first step, the LAM Crystal Philately competition has collected over 19.8 million valid structures, including 1 million on the OpenLAM convex hull, driving advancements in generative modeling and materials science applications.
2501.16360
Momentum Contrastive Learning with Enhanced Negative Sampling and Hard Negative Filtering
cs.LG cs.AI
Contrastive learning has become pivotal in unsupervised representation learning, with frameworks like Momentum Contrast (MoCo) effectively utilizing large negative sample sets to extract discriminative features. However, traditional approaches often overlook the full potential of key embeddings and are susceptible to performance degradation from noisy negative samples in the memory bank. This study addresses these challenges by proposing an enhanced contrastive learning framework that incorporates two key innovations. First, we introduce a dual-view loss function, which ensures balanced optimization of both query and key embeddings, improving representation quality. Second, we develop a selective negative sampling strategy that emphasizes the most challenging negatives based on cosine similarity, mitigating the impact of noise and enhancing feature discrimination. Extensive experiments demonstrate that our framework achieves superior performance on downstream tasks, delivering robust and well-structured representations. These results highlight the potential of optimized contrastive mechanisms to advance unsupervised learning and extend its applicability across domains such as computer vision and natural language processing
2501.16361
Large Language Models Meet Graph Neural Networks for Text-Numeric Graph Reasoning
cs.LG cs.AI
In real-world scientific discovery, human beings always make use of the accumulated prior knowledge with imagination pick select one or a few most promising hypotheses from large and noisy data analysis results. In this study, we introduce a new type of graph structure, the text-numeric graph (TNG), which is defined as graph entities and associations have both text-attributed information and numeric information. The TNG is an ideal data structure model for novel scientific discovery via graph reasoning because it integrates human-understandable textual annotations or prior knowledge, with numeric values that represent the observed or activation levels of graph entities or associations in different samples. Together both the textual information and numeric values determine the importance of graph entities and associations in graph reasoning for novel scientific knowledge discovery. We further propose integrating large language models (LLMs) and graph neural networks (GNNs) to analyze the TNGs for graph understanding and reasoning. To demonstrate the utility, we generated the text-omic(numeric) signaling graphs (TOSG), as one type of TNGs, in which all graphs have the same entities, associations and annotations, but have sample-specific entity numeric (omic) values using single cell RNAseq (scRNAseq) datasets of different diseases. We proposed joint LLM-GNN models for key entity mining and signaling pathway mining on the TOSGs. The evaluation results showed the LLM-GNN and TNGs models significantly improve classification accuracy and network inference. In conclusion, the TNGs and joint LLM-GNN models are important approaches for scientific discovery.
2501.16362
A novel Trunk Branch-net PINN for flow and heat transfer prediction in porous medium
cs.LG physics.flu-dyn
A novel Trunk-Branch (TB)-net physics-informed neural network (PINN) architecture is developed, which is a PINN-based method incorporating trunk and branch nets to capture both global and local features. The aim is to solve four main classes of problems: forward flow problem, forward heat transfer problem, inverse heat transfer problem, and transfer learning problem within the porous medium, which are notoriously complex that could not be handled by origin PINN. In the proposed TB-net PINN architecture, a Fully-connected Neural Network (FNN) is used as the trunk net, followed by separated FNNs as the branch nets with respect to outputs, and automatic differentiation is performed for partial derivatives of outputs with respect to inputs by considering various physical loss. The effectiveness and flexibility of the novel TB-net PINN architecture is demonstrated through a collection of forward problems, and transfer learning validates the feasibility of resource reuse. Combining with the superiority over traditional numerical methods in solving inverse problems, the proposed TB-net PINN shows its great potential for practical engineering applications.
2501.16364
Multivariate Time Series Anomaly Detection by Capturing Coarse-Grained Intra- and Inter-Variate Dependencies
cs.LG cs.AI
Multivariate time series anomaly detection is essential for failure management in web application operations, as it directly influences the effectiveness and timeliness of implementing remedial or preventive measures. This task is often framed as a semi-supervised learning problem, where only normal data are available for model training, primarily due to the labor-intensive nature of data labeling and the scarcity of anomalous data. Existing semi-supervised methods often detect anomalies by capturing intra-variate temporal dependencies and/or inter-variate relationships to learn normal patterns, flagging timestamps that deviate from these patterns as anomalies. However, these approaches often fail to capture salient intra-variate temporal and inter-variate dependencies in time series due to their focus on excessively fine granularity, leading to suboptimal performance. In this study, we introduce MtsCID, a novel semi-supervised multivariate time series anomaly detection method. MtsCID employs a dual network architecture: one network operates on the attention maps of multi-scale intra-variate patches for coarse-grained temporal dependency learning, while the other works on variates to capture coarse-grained inter-variate relationships through convolution and interaction with sinusoidal prototypes. This design enhances the ability to capture the patterns from both intra-variate temporal dependencies and inter-variate relationships, resulting in improved performance. Extensive experiments across seven widely used datasets demonstrate that MtsCID achieves performance comparable or superior to state-of-the-art benchmark methods.
2501.16365
CAND: Cross-Domain Ambiguity Inference for Early Detecting Nuanced Illness Deterioration
cs.LG cs.AI
Early detection of patient deterioration is essential for timely treatment, with vital signs like heart rates being key health indicators. Existing methods tend to solely analyze vital sign waveforms, ignoring transition relationships of waveforms within each vital sign and the correlation strengths among various vital signs. Such studies often overlook nuanced illness deterioration, which is the early sign of worsening health but is difficult to detect. In this paper, we introduce CAND, a novel method that organizes the transition relationships and the correlations within and among vital signs as domain-specific and cross-domain knowledge. CAND jointly models these knowledge in a unified representation space, considerably enhancing the early detection of nuanced illness deterioration. In addition, CAND integrates a Bayesian inference method that utilizes augmented knowledge from domain-specific and cross-domain knowledge to address the ambiguities in correlation strengths. With this architecture, the correlation strengths can be effectively inferred to guide joint modeling and enhance representations of vital signs. This allows a more holistic and accurate interpretation of patient health. Our experiments on a real-world ICU dataset demonstrate that CAND significantly outperforms existing methods in both effectiveness and earliness in detecting nuanced illness deterioration. Moreover, we conduct a case study for the interpretable detection process to showcase the practicality of CAND.
2501.16368
Foundation Models for CPS-IoT: Opportunities and Challenges
cs.LG cs.AI cs.SY eess.SY
Methods from machine learning (ML) have transformed the implementation of Perception-Cognition-Communication-Action loops in Cyber-Physical Systems (CPS) and the Internet of Things (IoT), replacing mechanistic and basic statistical models with those derived from data. However, the first generation of ML approaches, which depend on supervised learning with annotated data to create task-specific models, faces significant limitations in scaling to the diverse sensor modalities, deployment configurations, application tasks, and operating dynamics characterizing real-world CPS-IoT systems. The success of task-agnostic foundation models (FMs), including multimodal large language models (LLMs), in addressing similar challenges across natural language, computer vision, and human speech has generated considerable enthusiasm for and exploration of FMs and LLMs as flexible building blocks in CPS-IoT analytics pipelines, promising to reduce the need for costly task-specific engineering. Nonetheless, a significant gap persists between the current capabilities of FMs and LLMs in the CPS-IoT domain and the requirements they must meet to be viable for CPS-IoT applications. In this paper, we analyze and characterize this gap through a thorough examination of the state of the art and our research, which extends beyond it in various dimensions. Based on the results of our analysis and research, we identify essential desiderata that CPS-IoT domain-specific FMs and LLMs must satisfy to bridge this gap. We also propose actions by CPS-IoT researchers to collaborate in developing key community resources necessary for establishing FMs and LLMs as foundational tools for the next generation of CPS-IoT systems.
2501.16369
Blockchain-based Crowdsourced Deep Reinforcement Learning as a Service
cs.LG cs.AI
Deep Reinforcement Learning (DRL) has emerged as a powerful paradigm for solving complex problems. However, its full potential remains inaccessible to a broader audience due to its complexity, which requires expertise in training and designing DRL solutions, high computational capabilities, and sometimes access to pre-trained models. This necessitates the need for hassle-free services that increase the availability of DRL solutions to a variety of users. To enhance the accessibility to DRL services, this paper proposes a novel blockchain-based crowdsourced DRL as a Service (DRLaaS) framework. The framework provides DRL-related services to users, covering two types of tasks: DRL training and model sharing. Through crowdsourcing, users could benefit from the expertise and computational capabilities of workers to train DRL solutions. Model sharing could help users gain access to pre-trained models, shared by workers in return for incentives, which can help train new DRL solutions using methods in knowledge transfer. The DRLaaS framework is built on top of a Consortium Blockchain to enable traceable and autonomous execution. Smart Contracts are designed to manage worker and model allocation, which are stored using the InterPlanetary File System (IPFS) to ensure tamper-proof data distribution. The framework is tested on several DRL applications, proving its efficacy.
2501.16370
Advanced Physics-Informed Neural Network with Residuals for Solving Complex Integral Equations
cs.LG cs.AI cs.NA cs.NE math.NA
In this paper, we present the Residual Integral Solver Network (RISN), a novel neural network architecture designed to solve a wide range of integral and integro-differential equations, including one-dimensional, multi-dimensional, ordinary and partial integro-differential, systems, and fractional types. RISN integrates residual connections with high-accurate numerical methods such as Gaussian quadrature and fractional derivative operational matrices, enabling it to achieve higher accuracy and stability than traditional Physics-Informed Neural Networks (PINN). The residual connections help mitigate vanishing gradient issues, allowing RISN to handle deeper networks and more complex kernels, particularly in multi-dimensional problems. Through extensive experiments, we demonstrate that RISN consistently outperforms PINN, achieving significantly lower Mean Absolute Errors (MAE) across various types of equations. The results highlight RISN's robustness and efficiency in solving challenging integral and integro-differential problems, making it a valuable tool for real-world applications where traditional methods often struggle.
2501.16371
Which Optimizer Works Best for Physics-Informed Neural Networks and Kolmogorov-Arnold Networks?
cs.LG cs.AI math.OC
Physics-Informed Neural Networks (PINNs) have revolutionized the computation of PDE solutions by integrating partial differential equations (PDEs) into the neural network's training process as soft constraints, becoming an important component of the scientific machine learning (SciML) ecosystem. In its current implementation, PINNs are mainly optimized using first-order methods like Adam, as well as quasi-Newton methods such as BFGS and its low-memory variant, L-BFGS. However, these optimizers often struggle with highly non-linear and non-convex loss landscapes, leading to challenges such as slow convergence, local minima entrapment, and (non)degenerate saddle points. In this study, we investigate the performance of Self-Scaled Broyden (SSBroyden) methods and other advanced quasi-Newton schemes, including BFGS and L-BFGS with different line search strategies approaches. These methods dynamically rescale updates based on historical gradient information, thus enhancing training efficiency and accuracy. We systematically compare these optimizers on key challenging linear, stiff, multi-scale and non-linear PDEs benchmarks, including the Burgers, Allen-Cahn, Kuramoto-Sivashinsky, and Ginzburg-Landau equations, and extend our study to Physics-Informed Kolmogorov-Arnold Networks (PIKANs) representation. Our findings provide insights into the effectiveness of second-order optimization strategies in improving the convergence and accurate generalization of PINNs for complex PDEs by orders of magnitude compared to the state-of-the-art.
2501.16372
Low-Rank Adapters Meet Neural Architecture Search for LLM Compression
cs.LG cs.AI cs.CL
The rapid expansion of Large Language Models (LLMs) has posed significant challenges regarding the computational resources required for fine-tuning and deployment. Recent advancements in low-rank adapters have demonstrated their efficacy in parameter-efficient fine-tuning (PEFT) of these models. This retrospective paper comprehensively discusses innovative approaches that synergize low-rank representations with Neural Architecture Search (NAS) techniques, particularly weight-sharing super-networks. Robust solutions for compressing and fine-tuning large pre-trained models are developed by integrating these methodologies. Our analysis highlights the potential of these combined strategies to democratize the use of LLMs, making them more accessible for deployment in resource-constrained environments. The resulting models exhibit reduced memory footprints and faster inference times, paving the way for more practical and scalable applications of LLMs. Models and code are available at https://github.com/IntelLabs/Hardware-Aware-Automated-Machine-Learning.
2501.16373
Unveiling Discrete Clues: Superior Healthcare Predictions for Rare Diseases
cs.LG cs.AI cs.CE
Accurate healthcare prediction is essential for improving patient outcomes. Existing work primarily leverages advanced frameworks like attention or graph networks to capture the intricate collaborative (CO) signals in electronic health records. However, prediction for rare diseases remains challenging due to limited co-occurrence and inadequately tailored approaches. To address this issue, this paper proposes UDC, a novel method that unveils discrete clues to bridge consistent textual knowledge and CO signals within a unified semantic space, thereby enriching the representation semantics of rare diseases. Specifically, we focus on addressing two key sub-problems: (1) acquiring distinguishable discrete encodings for precise disease representation and (2) achieving semantic alignment between textual knowledge and the CO signals at the code level. For the first sub-problem, we refine the standard vector quantized process to include condition awareness. Additionally, we develop an advanced contrastive approach in the decoding stage, leveraging synthetic and mixed-domain targets as hard negatives to enrich the perceptibility of the reconstructed representation for downstream tasks. For the second sub-problem, we introduce a novel codebook update strategy using co-teacher distillation. This approach facilitates bidirectional supervision between textual knowledge and CO signals, thereby aligning semantically equivalent information in a shared discrete latent space. Extensive experiments on three datasets demonstrate our superiority.
2501.16374
SAFR: Neuron Redistribution for Interpretability
cs.LG cs.AI
Superposition refers to encoding representations of multiple features within a single neuron, which is common in deep neural networks. This property allows neurons to combine and represent multiple features, enabling the model to capture intricate information and handle complex tasks. Despite promising performance, the model's interpretability has been diminished. This paper presents a novel approach to enhance model interpretability by regularizing feature superposition. We introduce SAFR, which simply applies regularizations to the loss function to promote monosemantic representations for important tokens while encouraging polysemanticity for correlated token pairs, where important tokens and correlated token pairs are identified via VMASK and attention weights respectively. We evaluate SAFR with a transformer model on two classification tasks. Experiments demonstrate the effectiveness of SAFR in improving model interpretability without compromising prediction performance. Besides, SAFR provides explanations by visualizing the neuron allocation within the intermediate layers.
2501.16375
On Storage Neural Network Augmented Approximate Nearest Neighbor Search
cs.LG cs.AI cs.IR
Large-scale approximate nearest neighbor search (ANN) has been gaining attention along with the latest machine learning researches employing ANNs. If the data is too large to fit in memory, it is necessary to search for the most similar vectors to a given query vector from the data stored in storage devices, not from that in memory. The storage device such as NAND flash memory has larger capacity than the memory device such as DRAM, but they also have larger latency to read data. Therefore, ANN methods for storage require completely different approaches from conventional in-memory ANN methods. Since the approximation that the time required for search is determined only by the amount of data fetched from storage holds under reasonable assumptions, our goal is to minimize it while maximizing recall. For partitioning-based ANNs, vectors are partitioned into clusters in the index building phase. In the search phase, some of the clusters are chosen, the vectors in the chosen clusters are fetched from storage, and the nearest vector is retrieved from the fetched vectors. Thus, the key point is to accurately select the clusters containing the ground truth nearest neighbor vectors. We accomplish this by proposing a method to predict the correct clusters by means of a neural network that is gradually refined by alternating supervised learning and duplicated cluster assignment. Compared to state-of-the-art SPANN and an exhaustive method using k-means clustering and linear search, the proposed method achieves 90% recall on SIFT1M with 80% and 58% less data fetched from storage, respectively.
2501.16376
HWPQ: Hessian-free Weight Pruning-Quantization For LLM Compression And Acceleration
cs.LG cs.AI
Large Language Models (LLMs) have achieved remarkable success across numerous domains. However, the high time complexity of existing pruning and quantization methods significantly hinders their effective deployment on resource-constrained consumer or edge devices. In this study, we propose a novel Hessian-free Weight Pruning-Quantization (HWPQ) method. HWPQ eliminates the need for computationally intensive Hessian matrix calculations by introducing a contribution-based weight metric, which evaluates the importance of weights without relying on second-order derivatives. Additionally, we employ the Exponentially Weighted Moving Average (EWMA) technique to bypass weight sorting, enabling the selection of weights that contribute most to LLM accuracy and further reducing time complexity. Our approach is extended to support 2:4 structured sparsity pruning, facilitating efficient execution on modern hardware accelerators. Experimental results demonstrate that HWPQ significantly enhances the compression performance of LLaMA2. Compared to state-of-the-art quantization and pruning frameworks, HWPQ achieves average speedups of 5.97x (up to 20.75x) in quantization time and 12.29x (up to 56.02x) in pruning time, while largely preserving model accuracy. Furthermore, we observe a 1.50x inference speedup compared to the baseline.
2501.16377
Optimal Signal Decomposition-based Multi-Stage Learning for Battery Health Estimation
cs.LG cs.AI
Battery health estimation is fundamental to ensure battery safety and reduce cost. However, achieving accurate estimation has been challenging due to the batteries' complex nonlinear aging patterns and capacity regeneration phenomena. In this paper, we propose OSL, an optimal signal decomposition-based multi-stage machine learning for battery health estimation. OSL treats battery signals optimally. It uses optimized variational mode decomposition to extract decomposed signals capturing different frequency bands of the original battery signals. It also incorporates a multi-stage learning process to analyze both spatial and temporal battery features effectively. An experimental study is conducted with a public battery aging dataset. OSL demonstrates exceptional performance with a mean error of just 0.26%. It significantly outperforms comparison algorithms, both those without and those with suboptimal signal decomposition and analysis. OSL considers practical battery challenges and can be integrated into real-world battery management systems, offering a good impact on battery monitoring and optimization.
2501.16378
Internal Activation Revision: Safeguarding Vision Language Models Without Parameter Update
cs.LG cs.AI cs.CL cs.CV
Vision-language models (VLMs) demonstrate strong multimodal capabilities but have been found to be more susceptible to generating harmful content compared to their backbone large language models (LLMs). Our investigation reveals that the integration of images significantly shifts the model's internal activations during the forward pass, diverging from those triggered by textual input. Moreover, the safety alignments of LLMs embedded within VLMs are not sufficiently robust to handle the activations discrepancies, making the models vulnerable to even the simplest jailbreaking attacks. To address this issue, we propose an \textbf{internal activation revision} approach that efficiently revises activations during generation, steering the model toward safer outputs. Our framework incorporates revisions at both the layer and head levels, offering control over the model's generation at varying levels of granularity. In addition, we explore three strategies for constructing positive and negative samples and two approaches for extracting revision vectors, resulting in different variants of our method. Comprehensive experiments demonstrate that the internal activation revision method significantly improves the safety of widely used VLMs, reducing attack success rates by an average of 48.94\%, 34.34\%, 43.92\%, and 52.98\% on SafeBench, Safe-Unsafe, Unsafe, and MM-SafetyBench, respectively, while minimally impacting model helpfulness.
2501.16379
FedAGHN: Personalized Federated Learning with Attentive Graph HyperNetworks
cs.LG cs.AI
Personalized Federated Learning (PFL) aims to address the statistical heterogeneity of data across clients by learning the personalized model for each client. Among various PFL approaches, the personalized aggregation-based approach conducts parameter aggregation in the server-side aggregation phase to generate personalized models, and focuses on learning appropriate collaborative relationships among clients for aggregation. However, the collaborative relationships vary in different scenarios and even at different stages of the FL process. To this end, we propose Personalized Federated Learning with Attentive Graph HyperNetworks (FedAGHN), which employs Attentive Graph HyperNetworks (AGHNs) to dynamically capture fine-grained collaborative relationships and generate client-specific personalized initial models. Specifically, AGHNs empower graphs to explicitly model the client-specific collaborative relationships, construct collaboration graphs, and introduce tunable attentive mechanism to derive the collaboration weights, so that the personalized initial models can be obtained by aggregating parameters over the collaboration graphs. Extensive experiments can demonstrate the superiority of FedAGHN. Moreover, a series of visualizations are presented to explore the effectiveness of collaboration graphs learned by FedAGHN.
2501.16380
UDiTQC: U-Net-Style Diffusion Transformer for Quantum Circuit Synthesis
cs.LG cs.AI quant-ph
Quantum computing is a transformative technology with wide-ranging applications, and efficient quantum circuit generation is crucial for unlocking its full potential. Current diffusion model approaches based on U-Net architectures, while promising, encounter challenges related to computational efficiency and modeling global context. To address these issues, we propose UDiT,a novel U-Net-style Diffusion Transformer architecture, which combines U-Net's strengths in multi-scale feature extraction with the Transformer's ability to model global context. We demonstrate the framework's effectiveness on two tasks: entanglement generation and unitary compilation, where UDiTQC consistently outperforms existing methods. Additionally, our framework supports tasks such as masking and editing circuits to meet specific physical property requirements. This dual advancement, improving quantum circuit synthesis and refining generative model architectures, marks a significant milestone in the convergence of quantum computing and machine learning research.
2501.16381
Reduced-order modeling and classification of hydrodynamic pattern formation in gravure printing
cs.LG physics.flu-dyn
Hydrodynamic pattern formation phenomena in printing and coating processes are still not fully understood. However, fundamental understanding is essential to achieve high-quality printed products and to tune printed patterns according to the needs of a specific application like printed electronics, graphical printing, or biomedical printing. The aim of the paper is to develop an automated pattern classification algorithm based on methods from supervised machine learning and reduced-order modeling. We use the HYPA-p dataset, a large image dataset of gravure-printed images, which shows various types of hydrodynamic pattern formation phenomena. It enables the correlation of printing process parameters and resulting printed patterns for the first time. 26880 images of the HYPA-p dataset have been labeled by a human observer as dot patterns, mixed patterns, or finger patterns; 864000 images (97%) are unlabeled. A singular value decomposition (SVD) is used to find the modes of the labeled images and to reduce the dimensionality of the full dataset by truncation and projection. Selected machine learning classification techniques are trained on the reduced-order data. We investigate the effect of several factors, including classifier choice, whether or not fast Fourier transform (FFT) is used to preprocess the labeled images, data balancing, and data normalization. The best performing model is a k-nearest neighbor (kNN) classifier trained on unbalanced, FFT-transformed data with a test error of 3%, which outperforms a human observer by 7%. Data balancing slightly increases the test error of the kNN-model to 5%, but also increases the recall of the mixed class from 90% to 94%. Finally, we demonstrate how the trained models can be used to predict the pattern class of unlabeled images and how the predictions can be correlated to the printing process parameters, in the form of regime maps.
2501.16382
GraPPI: A Retrieve-Divide-Solve GraphRAG Framework for Large-scale Protein-protein Interaction Exploration
q-bio.QM cs.AI cs.LG
Drug discovery (DD) has tremendously contributed to maintaining and improving public health. Hypothesizing that inhibiting protein misfolding can slow disease progression, researchers focus on target identification (Target ID) to find protein structures for drug binding. While Large Language Models (LLMs) and Retrieval-Augmented Generation (RAG) frameworks have accelerated drug discovery, integrating models into cohesive workflows remains challenging. We conducted a user study with drug discovery researchers to identify the applicability of LLMs and RAGs in Target ID. We identified two main findings: 1) an LLM should provide multiple Protein-Protein Interactions (PPIs) based on an initial protein and protein candidates that have a therapeutic impact; 2) the model must provide the PPI and relevant explanations for better understanding. Based on these observations, we identified three limitations in previous approaches for Target ID: 1) semantic ambiguity, 2) lack of explainability, and 3) short retrieval units. To address these issues, we propose GraPPI, a large-scale knowledge graph (KG)-based retrieve-divide-solve agent pipeline RAG framework to support large-scale PPI signaling pathway exploration in understanding therapeutic impacts by decomposing the analysis of entire PPI pathways into sub-tasks focused on the analysis of PPI edges.
2501.16383
RotateKV: Accurate and Robust 2-Bit KV Cache Quantization for LLMs via Outlier-Aware Adaptive Rotations
cs.LG cs.AI cs.CL
Key-Value (KV) cache facilitates efficient large language models (LLMs) inference by avoiding recomputation of past KVs. As the batch size and context length increase, the oversized KV caches become a significant memory bottleneck, highlighting the need for efficient compression. Existing KV quantization rely on fine-grained quantization or the retention of a significant portion of high bit-widths caches, both of which compromise compression ratio and often fail to maintain robustness at extremely low average bit-widths. In this work, we explore the potential of rotation technique for 2-bit KV quantization and propose RotateKV, which achieves accurate and robust performance through the following innovations: (i) Outlier-Aware Rotation, which utilizes channel-reordering to adapt the rotations to varying channel-wise outlier distributions without sacrificing the computational efficiency of the fast Walsh-Hadamard transform (FWHT); (ii) Pre-RoPE Grouped-Head Rotation, which mitigates the impact of rotary position embedding (RoPE) on proposed outlier-aware rotation and further smooths outliers across heads; (iii) Attention-Sink-Aware Quantization, which leverages the massive activations to precisely identify and protect attention sinks. RotateKV achieves less than 0.3 perplexity (PPL) degradation with 2-bit quantization on WikiText-2 using LLaMA-2-13B, maintains strong CoT reasoning and long-context capabilities, with less than 1.7\% degradation on GSM8K, outperforming existing methods even at lower average bit-widths. RotateKV also showcases a 3.97x reduction in peak memory usage, supports 5.75x larger batch sizes, and achieves a 2.32x speedup in decoding stage.
2501.16384
MambaTron: Efficient Cross-Modal Point Cloud Enhancement using Aggregate Selective State Space Modeling
eess.SP cs.LG
Point cloud enhancement is the process of generating a high-quality point cloud from an incomplete input. This is done by filling in the missing details from a reference like the ground truth via regression, for example. In addition to unimodal image and point cloud reconstruction, we focus on the task of view-guided point cloud completion, where we gather the missing information from an image, which represents a view of the point cloud and use it to generate the output point cloud. With the recent research efforts surrounding state-space models, originally in natural language processing and now in 2D and 3D vision, Mamba has shown promising results as an efficient alternative to the self-attention mechanism. However, there is limited research towards employing Mamba for cross-attention between the image and the input point cloud, which is crucial in multi-modal problems. In this paper, we introduce MambaTron, a Mamba-Transformer cell that serves as a building block for our network which is capable of unimodal and cross-modal reconstruction which includes view-guided point cloud completion.We explore the benefits of Mamba's long-sequence efficiency coupled with the Transformer's excellent analytical capabilities through MambaTron. This approach is one of the first attempts to implement a Mamba-based analogue of cross-attention, especially in computer vision. Our model demonstrates a degree of performance comparable to the current state-of-the-art techniques while using a fraction of the computation resources.
2501.16385
FBQuant: FeedBack Quantization for Large Language Models
cs.LG cs.CL
Deploying Large Language Models (LLMs) on edge devices is increasingly important, as it eliminates reliance on network connections, reduces expensive API calls, and enhances user privacy. However, on-device deployment is challenging due to the limited computational resources of edge devices. In particular, the key bottleneck stems from memory bandwidth constraints related to weight loading. Weight-only quantization effectively reduces memory access, yet often induces significant accuracy degradation. Recent efforts to incorporate sub-branches have shown promise for mitigating quantization errors, but these methods either lack robust optimization strategies or rely on suboptimal objectives. To address these gaps, we propose FeedBack Quantization (FBQuant), a novel approach inspired by negative feedback mechanisms in automatic control. FBQuant inherently ensures that the reconstructed weights remain bounded by the quantization process, thereby reducing the risk of overfitting. To further offset the additional latency introduced by sub-branches, we develop an efficient CUDA kernel that decreases 60\% of extra inference time. Comprehensive experiments demonstrate the efficiency and effectiveness of FBQuant across various LLMs. Notably, for 3-bit Llama2-7B, FBQuant improves zero-shot accuracy by 1.2\%.
2501.16386
ILETIA: An AI-enhanced method for individualized trigger-oocyte pickup interval estimation of progestin-primed ovarian stimulation protocol
q-bio.QM cs.LG
In vitro fertilization-embryo transfer (IVF-ET) stands as one of the most prevalent treatments for infertility. During an IVF-ET cycle, the time interval between trigger shot and oocyte pickup (OPU) is a pivotal period for follicular maturation, which determines mature oocytes yields and impacts the success of subsequent procedures. However, accurately predicting this interval is severely hindered by the variability of clinicians'experience that often leads to suboptimal oocyte retrieval rate. To address this challenge, we propose ILETIA, the first machine learning-based method that could predict the optimal trigger-OPU interval for patients receiving progestin-primed ovarian stimulation (PPOS) protocol. Specifically, ILETIA leverages a Transformer to learn representations from clinical tabular data, and then employs gradient-boosted trees for interval prediction. For model training and evaluating, we compiled a dataset PPOS-DS of nearly ten thousand patients receiving PPOS protocol, the largest such dataset to our knowledge. Experimental results demonstrate that our method achieves strong performance (AUROC = 0.889), outperforming both clinicians and other widely used computational models. Moreover, ILETIA also supports premature ovulation risk prediction in a specific OPU time (AUROC = 0.838). Collectively, by enabling more precise and individualized decisions, ILETIA has the potential to improve clinical outcomes and lay the foundation for future IVF-ET research.
2501.16388
Development and Validation of a Dynamic Kidney Failure Prediction Model based on Deep Learning: A Real-World Study with External Validation
cs.LG stat.AP
Background: Chronic kidney disease (CKD), a progressive disease with high morbidity and mortality, has become a significant global public health problem. At present, most of the models used for predicting the progression of CKD are static models. We aim to develop a dynamic kidney failure prediction model based on deep learning (KFDeep) for CKD patients, utilizing all available data on common clinical indicators from real-world Electronic Health Records (EHRs) to provide real-time predictions. Findings: A retrospective cohort of 4,587 patients from EHRs of Yinzhou, China, is used as the development dataset (2,752 patients for training, 917 patients for validation) and internal validation dataset (917 patients), while a prospective cohort of 934 patients from the Peking University First Hospital CKD cohort (PKUFH cohort) is used as the external validation dataset. The AUROC of the KFDeep model reaches 0.946 (95\% CI: 0.922-0.970) on the internal validation dataset and 0.805 (95\% CI: 0.763-0.847) on the external validation dataset, both surpassing existing models. The KFDeep model demonstrates stable performance in simulated dynamic scenarios, with the AUROC progressively increasing over time. Both the calibration curve and decision curve analyses confirm that the model is unbiased and safe for practical use, while the SHAP analysis and hidden layer clustering results align with established medical knowledge. Interpretation: The KFDeep model built from real-world EHRs enhances the prediction accuracy of kidney failure without increasing clinical examination costs and can be easily integrated into existing hospital systems, providing physicians with a continuously updated decision-support tool due to its dynamic design.
2501.16389
Bridging the Sim2Real Gap: Vision Encoder Pre-Training for Visuomotor Policy Transfer
cs.RO cs.CV
Simulation offers a scalable and efficient alternative to real-world data collection for learning visuomotor robotic policies. However, the simulation-to-reality, or "Sim2Real" distribution shift -- introduced by employing simulation-trained policies in real-world environments -- frequently prevents successful policy transfer. This study explores the potential of using large-scale pre-training of vision encoders to address the Sim2Real gap. We examine a diverse collection of encoders, evaluating their ability to (1) extract features necessary for robot control while (2) remaining invariant to task-irrelevant environmental variations. We quantitatively measure the encoder's feature extraction capabilities through linear probing and its domain invariance by computing distances between simulation and real-world embedding centroids. Additional qualitative insights are provided through t-SNE plots and GradCAM saliency maps. Findings suggest that encoders pre-trained on manipulation-specific datasets generally outperform those trained on generic datasets in bridging the Sim2Real gap. https://github.com/yyardi/Bridging-the-Sim2Real-Gap
2501.16391
Leveraging Induced Transferable Binding Principles for Associative Prediction of Novel Drug-Target Interactions
cs.LG cs.AI q-bio.BM
Significant differences in protein structures hinder the generalization of existing drug-target interaction (DTI) models, which often rely heavily on pre-learned binding principles or detailed annotations. In contrast, BioBridge designs an Inductive-Associative pipeline inspired by the workflow of scientists who base their accumulated expertise on drawing insights into novel drug-target pairs from weakly related references. BioBridge predicts novel drug-target interactions using limited sequence data, incorporating multi-level encoders with adversarial training to accumulate transferable binding principles. On these principles basis, BioBridge employs a dynamic prototype meta-learning framework to associate insights from weakly related annotations, enabling robust predictions for previously unseen drug-target pairs. Extensive experiments demonstrate that BioBridge surpasses existing models, especially for unseen proteins. Notably, when only homologous protein binding data is available, BioBridge proves effective for virtual screening of the epidermal growth factor receptor and adenosine receptor, underscoring its potential in drug discovery.
2501.16392
HMCGeo: IP Region Prediction Based on Hierarchical Multi-label Classification
cs.LG
Fine-grained IP geolocation plays a critical role in applications such as location-based services and cybersecurity. Most existing fine-grained IP geolocation methods are regression-based; however, due to noise in the input data, these methods typically encounter kilometer-level prediction errors and provide incorrect region information for users. To address this issue, this paper proposes a novel hierarchical multi-label classification framework for IP region prediction, named HMCGeo. This framework treats IP geolocation as a hierarchical multi-label classification problem and employs residual connection-based feature extraction and attention prediction units to predict the target host region across multiple geographical granularities. Furthermore, we introduce probabilistic classification loss during training, combining it with hierarchical cross-entropy loss to form a composite loss function. This approach optimizes predictions by utilizing hierarchical constraints between regions at different granularities. IP region prediction experiments on the New York, Los Angeles, and Shanghai datasets demonstrate that HMCGeo achieves superior performance across all geographical granularities, significantly outperforming existing IP geolocation methods.
2501.16393
Improving Network Threat Detection by Knowledge Graph, Large Language Model, and Imbalanced Learning
cs.LG cs.CR stat.ML
Network threat detection has been challenging due to the complexities of attack activities and the limitation of historical threat data to learn from. To help enhance the existing practices of using analytics, machine learning, and artificial intelligence methods to detect the network threats, we propose an integrated modelling framework, where Knowledge Graph is used to analyze the users' activity patterns, Imbalanced Learning techniques are used to prune and weigh Knowledge Graph, and LLM is used to retrieve and interpret the users' activities from Knowledge Graph. The proposed framework is applied to Agile Threat Detection through Online Sequential Learning. The preliminary results show the improved threat capture rate by 3%-4% and the increased interpretabilities of risk predictions based on the users' activities.
2501.16394
Transformer^-1: Input-Adaptive Computation for Resource-Constrained Deployment
cs.LG
Addressing the resource waste caused by fixed computation paradigms in deep learning models under dynamic scenarios, this paper proposes a Transformer$^{-1}$ architecture based on the principle of deep adaptivity. This architecture achieves dynamic matching between input features and computational resources by establishing a joint optimization model for complexity and computation. Our core contributions include: (1) designing a two-layer control mechanism, composed of a complexity predictor and a reinforcement learning policy network, enabling end-to-end optimization of computation paths; (2) deriving a lower bound theory for dynamic computation, proving the system's theoretical reach to optimal efficiency; and (3) proposing a layer folding technique and a CUDA Graph pre-compilation scheme, overcoming the engineering bottlenecks of dynamic architectures. In the ImageNet-1K benchmark test, our method reduces FLOPs by 42.7\% and peak memory usage by 34.1\% compared to the standard Transformer, while maintaining comparable accuracy ($\pm$0.3\%). Furthermore, we conducted practical deployment on the Jetson AGX Xavier platform, verifying the effectiveness and practical value of this method in resource-constrained environments. To further validate the generality of the method, we also conducted experiments on several natural language processing tasks and achieved significant improvements in resource efficiency.
2501.16396
TopoNets: High Performing Vision and Language Models with Brain-Like Topography
cs.LG cs.NE q-bio.NC
Neurons in the brain are organized such that nearby cells tend to share similar functions. AI models lack this organization, and past efforts to introduce topography have often led to trade-offs between topography and task performance. In this work, we present TopoLoss, a new loss function that promotes spatially organized topographic representations in AI models without significantly sacrificing task performance. TopoLoss is highly adaptable and can be seamlessly integrated into the training of leading model architectures. We validate our method on both vision (ResNet-18, ResNet-50, ViT) and language models (GPT-Neo-125M, NanoGPT), collectively TopoNets. TopoNets are the highest-performing supervised topographic models to date, exhibiting brain-like properties such as localized feature processing, lower dimensionality, and increased efficiency. TopoNets also predict responses in the brain and replicate the key topographic signatures observed in the brain's visual and language cortices. Together, this work establishes a robust and generalizable framework for integrating topography into leading model architectures, advancing the development of high-performing models that more closely emulate the computational strategies of the human brain.
2501.16397
THOR: A Generic Energy Estimation Approach for On-Device Training
cs.LG
Battery-powered mobile devices (e.g., smartphones, AR/VR glasses, and various IoT devices) are increasingly being used for AI training due to their growing computational power and easy access to valuable, diverse, and real-time data. On-device training is highly energy-intensive, making accurate energy consumption estimation crucial for effective job scheduling and sustainable AI. However, the heterogeneity of devices and the complexity of models challenge the accuracy and generalizability of existing estimation methods. This paper proposes THOR, a generic approach for energy consumption estimation in deep neural network (DNN) training. First, we examine the layer-wise energy additivity property of DNNs and strategically partition the entire model into layers for fine-grained energy consumption profiling. Then, we fit Gaussian Process (GP) models to learn from layer-wise energy consumption measurements and estimate a DNN's overall energy consumption based on its layer-wise energy additivity property. We conduct extensive experiments with various types of models across different real-world platforms. The results demonstrate that THOR has effectively reduced the Mean Absolute Percentage Error (MAPE) by up to 30%. Moreover, THOR is applied in guiding energy-aware pruning, successfully reducing energy consumption by 50%, thereby further demonstrating its generality and potential.
2501.16398
Visualizing the Local Atomic Environment Features of Machine Learning Interatomic Potential
cs.LG physics.atom-ph
This paper addresses the challenges of creating efficient and high-quality datasets for machine learning potential functions. We present a novel approach, termed DV-LAE (Difference Vectors based on Local Atomic Environments), which utilizes the properties of atomic local environments and employs histogram statistics to generate difference vectors. This technique facilitates dataset screening and optimization, effectively minimizing redundancy while maintaining data diversity. We have validated the optimized datasets in high-temperature and high-pressure hydrogen systems as well as the {\alpha}-Fe/H binary system, demonstrating a significant reduction in computational resource usage without compromising prediction accuracy. Additionally, our method has revealed new structures that emerge during simulations but were underrepresented in the initial training datasets. The redundancy in the datasets and the distribution of these new structures can be visually analyzed through the visualization of difference vectors. This approach enhances our understanding of the characteristics of these newly formed structures and their impact on physical processes.
2501.16399
Detecting clinician implicit biases in diagnoses using proximal causal inference
cs.LG stat.AP
Clinical decisions to treat and diagnose patients are affected by implicit biases formed by racism, ableism, sexism, and other stereotypes. These biases reflect broader systemic discrimination in healthcare and risk marginalizing already disadvantaged groups. Existing methods for measuring implicit biases require controlled randomized testing and only capture individual attitudes rather than outcomes. However, the "big-data" revolution has led to the availability of large observational medical datasets, like EHRs and biobanks, that provide the opportunity to investigate discrepancies in patient health outcomes. In this work, we propose a causal inference approach to detect the effect of clinician implicit biases on patient outcomes in large-scale medical data. Specifically, our method uses proximal mediation to disentangle pathway-specific effects of a patient's sociodemographic attribute on a clinician's diagnosis decision. We test our method on real-world data from the UK Biobank. Our work can serve as a tool that initiates conversation and brings awareness to unequal health outcomes caused by implicit biases.
2501.16403
Is Open Source the Future of AI? A Data-Driven Approach
cs.SE cs.AI cs.CL
Large Language Models (LLMs) have become central in academia and industry, raising concerns about privacy, transparency, and misuse. A key issue is the trustworthiness of proprietary models, with open-sourcing often proposed as a solution. However, open-sourcing presents challenges, including potential misuse, financial disincentives, and intellectual property concerns. Proprietary models, backed by private sector resources, are better positioned for return on investment. There are also other approaches that lie somewhere on the spectrum between completely open-source and proprietary. These can largely be categorised into open-source usage limitations protected by licensing, partially open-source (open weights) models, hybrid approaches where obsolete model versions are open-sourced, while competitive versions with market value remain proprietary. Currently, discussions on where on the spectrum future models should fall on remains unbacked and mostly opinionated where industry leaders are weighing in on the discussion. In this paper, we present a data-driven approach by compiling data on open-source development of LLMs, and their contributions in terms of improvements, modifications, and methods. Our goal is to avoid supporting either extreme but rather present data that will support future discussions both by industry experts as well as policy makers. Our findings indicate that open-source contributions can enhance model performance, with trends such as reduced model size and manageable accuracy loss. We also identify positive community engagement patterns and architectures that benefit most from open contributions.
2501.16404
DynaPrompt: Dynamic Test-Time Prompt Tuning
cs.LG cs.AI cs.CL
Test-time prompt tuning enhances zero-shot generalization of vision-language models but tends to ignore the relatedness among test samples during inference. Online test-time prompt tuning provides a simple way to leverage the information in previous test samples, albeit with the risk of prompt collapse due to error accumulation. To enhance test-time prompt tuning, we propose DynaPrompt, short for dynamic test-time prompt tuning, exploiting relevant data distribution information while reducing error accumulation. Built on an online prompt buffer, DynaPrompt adaptively selects and optimizes the relevant prompts for each test sample during tuning. Specifically, we introduce a dynamic prompt selection strategy based on two metrics: prediction entropy and probability difference. For unseen test data information, we develop dynamic prompt appending, which allows the buffer to append new prompts and delete the inactive ones. By doing so, the prompts are optimized to exploit beneficial information on specific test data, while alleviating error accumulation. Experiments on fourteen datasets demonstrate the effectiveness of dynamic test-time prompt tuning.
2501.16405
DepoRanker: A Web Tool to predict Klebsiella Depolymerases using Machine Learning
q-bio.GN cs.LG
Background: Phage therapy shows promise for treating antibiotic-resistant Klebsiella infections. Identifying phage depolymerases that target Klebsiella capsular polysaccharides is crucial, as these capsules contribute to biofilm formation and virulence. However, homology-based searches have limitations in novel depolymerase discovery. Objective: To develop a machine learning model for identifying and ranking potential phage depolymerases targeting Klebsiella. Methods: We developed DepoRanker, a machine learning algorithm to rank proteins by their likelihood of being depolymerases. The model was experimentally validated on 5 newly characterized proteins and compared to BLAST. Results: DepoRanker demonstrated superior performance to BLAST in identifying potential depolymerases. Experimental validation confirmed its predictive ability on novel proteins. Conclusions: DepoRanker provides an accurate and functional tool to expedite depolymerase discovery for phage therapy against Klebsiella. It is available as a webserver and open-source software. Availability: Webserver: https://deporanker.dcs.warwick.ac.uk/ Source code: https://github.com/wgrgwrght/deporanker
2501.16409
Classification of Mild Cognitive Impairment Based on Dynamic Functional Connectivity Using Spatio-Temporal Transformer
eess.IV cs.AI q-bio.NC
Dynamic functional connectivity (dFC) using resting-state functional magnetic resonance imaging (rs-fMRI) is an advanced technique for capturing the dynamic changes of neural activities, and can be very useful in the studies of brain diseases such as Alzheimer's disease (AD). Yet, existing studies have not fully leveraged the sequential information embedded within dFC that can potentially provide valuable information when identifying brain conditions. In this paper, we propose a novel framework that jointly learns the embedding of both spatial and temporal information within dFC based on the transformer architecture. Specifically, we first construct dFC networks from rs-fMRI data through a sliding window strategy. Then, we simultaneously employ a temporal block and a spatial block to capture higher-order representations of dynamic spatio-temporal dependencies, via mapping them into an efficient fused feature representation. To further enhance the robustness of these feature representations by reducing the dependency on labeled data, we also introduce a contrastive learning strategy to manipulate different brain states. Experimental results on 345 subjects with 570 scans from the Alzheimer's Disease Neuroimaging Initiative (ADNI) demonstrate the superiority of our proposed method for MCI (Mild Cognitive Impairment, the prodromal stage of AD) prediction, highlighting its potential for early identification of AD.
2501.16410
DynAlign: Unsupervised Dynamic Taxonomy Alignment for Cross-Domain Segmentation
cs.CV
Current unsupervised domain adaptation (UDA) methods for semantic segmentation typically assume identical class labels between the source and target domains. This assumption ignores the label-level domain gap, which is common in real-world scenarios, thus limiting their ability to identify finer-grained or novel categories without requiring extensive manual annotation. A promising direction to address this limitation lies in recent advancements in foundation models, which exhibit strong generalization abilities due to their rich prior knowledge. However, these models often struggle with domain-specific nuances and underrepresented fine-grained categories. To address these challenges, we introduce DynAlign, a framework that integrates UDA with foundation models to bridge both the image-level and label-level domain gaps. Our approach leverages prior semantic knowledge to align source categories with target categories that can be novel, more fine-grained, or named differently (e.g., vehicle to {car, truck, bus}). Foundation models are then employed for precise segmentation and category reassignment. To further enhance accuracy, we propose a knowledge fusion approach that dynamically adapts to varying scene contexts. DynAlign generates accurate predictions in a new target label space without requiring any manual annotations, allowing seamless adaptation to new taxonomies through either model retraining or direct inference. Experiments on the street scene semantic segmentation benchmarks GTA to Mapillary Vistas and GTA to IDD validate the effectiveness of our approach, achieving a significant improvement over existing methods. Our code will be publicly available.
2501.16411
PhysBench: Benchmarking and Enhancing Vision-Language Models for Physical World Understanding
cs.CV cs.AI cs.CL cs.LG cs.RO
Understanding the physical world is a fundamental challenge in embodied AI, critical for enabling agents to perform complex tasks and operate safely in real-world environments. While Vision-Language Models (VLMs) have shown great promise in reasoning and task planning for embodied agents, their ability to comprehend physical phenomena remains extremely limited. To close this gap, we introduce PhysBench, a comprehensive benchmark designed to evaluate VLMs' physical world understanding capability across a diverse set of tasks. PhysBench contains 10,002 entries of interleaved video-image-text data, categorized into four major domains: physical object properties, physical object relationships, physical scene understanding, and physics-based dynamics, further divided into 19 subclasses and 8 distinct capability dimensions. Our extensive experiments, conducted on 75 representative VLMs, reveal that while these models excel in common-sense reasoning, they struggle with understanding the physical world -- likely due to the absence of physical knowledge in their training data and the lack of embedded physical priors. To tackle the shortfall, we introduce PhysAgent, a novel framework that combines the generalization strengths of VLMs with the specialized expertise of vision models, significantly enhancing VLMs' physical understanding across a variety of tasks, including an 18.4\% improvement on GPT-4o. Furthermore, our results demonstrate that enhancing VLMs' physical world understanding capabilities can help embodied agents such as MOKA. We believe that PhysBench and PhysAgent offer valuable insights and contribute to bridging the gap between VLMs and physical world understanding.
2501.16443
Objects matter: object-centric world models improve reinforcement learning in visually complex environments
cs.LG cs.CV
Deep reinforcement learning has achieved remarkable success in learning control policies from pixels across a wide range of tasks, yet its application remains hindered by low sample efficiency, requiring significantly more environment interactions than humans to reach comparable performance. Model-based reinforcement learning (MBRL) offers a solution by leveraging learnt world models to generate simulated experience, thereby improving sample efficiency. However, in visually complex environments, small or dynamic elements can be critical for decision-making. Yet, traditional MBRL methods in pixel-based environments typically rely on auto-encoding with an $L_2$ loss, which is dominated by large areas and often fails to capture decision-relevant details. To address these limitations, we propose an object-centric MBRL pipeline, which integrates recent advances in computer vision to allow agents to focus on key decision-related elements. Our approach consists of four main steps: (1) annotating key objects related to rewards and goals with segmentation masks, (2) extracting object features using a pre-trained, frozen foundation vision model, (3) incorporating these object features with the raw observations to predict environmental dynamics, and (4) training the policy using imagined trajectories generated by this object-centric world model. Building on the efficient MBRL algorithm STORM, we call this pipeline OC-STORM. We demonstrate OC-STORM's practical value in overcoming the limitations of conventional MBRL approaches on both Atari games and the visually complex game Hollow Knight.
2501.16448
What is Harm? Baby Don't Hurt Me! On the Impossibility of Complete Harm Specification in AI Alignment
cs.AI cs.LG
"First, do no harm" faces a fundamental challenge in artificial intelligence: how can we specify what constitutes harm? While prior work treats harm specification as a technical hurdle to be overcome through better algorithms or more data, we argue this assumption is unsound. Drawing on information theory, we demonstrate that complete harm specification is fundamentally impossible for any system where harm is defined external to its specifications. This impossibility arises from an inescapable information-theoretic gap: the entropy of harm H(O) always exceeds the mutual information I(O;I) between ground truth harm O and a system's specifications I. We introduce two novel metrics: semantic entropy H(S) and the safety-capability ratio I(O;I)/H(O), to quantify these limitations. Through a progression of increasingly sophisticated specification attempts, we show why each approach must fail and why the resulting gaps are not mere engineering challenges but fundamental constraints akin to the halting problem. These results suggest a paradigm shift: rather than pursuing complete specifications, AI alignment research should focus on developing systems that can operate safely despite irreducible specification uncertainty.
2501.16450
360Brew: A Decoder-only Foundation Model for Personalized Ranking and Recommendation
cs.IR cs.AI
Ranking and recommendation systems are the foundation for numerous online experiences, ranging from search results to personalized content delivery. These systems have evolved into complex, multilayered architectures that leverage vast datasets and often incorporate thousands of predictive models. The maintenance and enhancement of these models is a labor intensive process that requires extensive feature engineering. This approach not only exacerbates technical debt but also hampers innovation in extending these systems to emerging problem domains. In this report, we present our research to address these challenges by utilizing a large foundation model with a textual interface for ranking and recommendation tasks. We illustrate several key advantages of our approach: (1) a single model can manage multiple predictive tasks involved in ranking and recommendation, (2) decoder models with textual interface due to their comprehension of reasoning capabilities, can generalize to new recommendation surfaces and out-of-domain problems, and (3) by employing natural language interfaces for task definitions and verbalizing member behaviors and their social connections, we eliminate the need for feature engineering and the maintenance of complex directed acyclic graphs of model dependencies. We introduce our research pre-production model, 360Brew V1.0, a 150B parameter, decoder-only model that has been trained and fine-tuned on LinkedIn's data and tasks. This model is capable of solving over 30 predictive tasks across various segments of the LinkedIn platform, achieving performance levels comparable to or exceeding those of current production systems based on offline metrics, without task-specific fine-tuning. Notably, each of these tasks is conventionally addressed by dedicated models that have been developed and maintained over multiple years by teams of a similar or larger size than our own.
2501.16453
Detecting Zero-Day Attacks in Digital Substations via In-Context Learning
cs.LG cs.AI
The occurrences of cyber attacks on the power grids have been increasing every year, with novel attack techniques emerging every year. In this paper, we address the critical challenge of detecting novel/zero-day attacks in digital substations that employ the IEC-61850 communication protocol. While many heuristic and machine learning (ML)-based methods have been proposed for attack detection in IEC-61850 digital substations, generalization to novel or zero-day attacks remains challenging. We propose an approach that leverages the in-context learning (ICL) capability of the transformer architecture, the fundamental building block of large language models. The ICL approach enables the model to detect zero-day attacks and learn from a few examples of that attack without explicit retraining. Our experiments on the IEC-61850 dataset demonstrate that the proposed method achieves more than $85\%$ detection accuracy on zero-day attacks while the existing state-of-the-art baselines fail. This work paves the way for building more secure and resilient digital substations of the future.
2501.16456
CoCoNUT: Structural Code Understanding does not fall out of a tree
cs.LG cs.SE
Large Language Models (LLMs) have shown impressive performance across a wide array of tasks involving both structured and unstructured textual data. Recent results on various benchmarks for code generation, repair, or completion suggest that certain models have programming abilities comparable to or even surpass humans. In this work, we demonstrate that high performance on such benchmarks does not correlate to humans' innate ability to understand structural control flow in code. To this end, we extract solutions from the HumanEval benchmark, which the relevant models perform strongly on, and trace their execution path using function calls sampled from the respective test set. Using this dataset, we investigate the ability of seven state-of-the-art LLMs to match the execution trace and find that, despite their ability to generate semantically identical code, they possess limited ability to trace execution paths, especially for longer traces and specific control structures. We find that even the top-performing model, Gemini, can fully and correctly generate only 47% of HumanEval task traces. Additionally, we introduce a subset for three key structures not contained in HumanEval: Recursion, Parallel Processing, and Object-Oriented Programming, including concepts like Inheritance and Polymorphism. Besides OOP, we show that none of the investigated models achieve an accuracy over 5% on the relevant traces. Aggregating these specialized parts with HumanEval tasks, we present CoCoNUT: Code Control Flow for Navigation Understanding and Testing, which measures a model's ability to trace execution of code upon relevant calls, including advanced structural components. We conclude that current LLMs need significant improvement to enhance code reasoning abilities. We hope our dataset helps researchers bridge this gap.
2501.16458
BiFold: Bimanual Cloth Folding with Language Guidance
cs.RO cs.CV
Cloth folding is a complex task due to the inevitable self-occlusions of clothes, their complicated dynamics, and the disparate materials, geometries, and textures that garments can have. In this work, we learn folding actions conditioned on text commands. Translating high-level, abstract instructions into precise robotic actions requires sophisticated language understanding and manipulation capabilities. To do that, we leverage a pre-trained vision-language model and repurpose it to predict manipulation actions. Our model, BiFold, can take context into account and achieves state-of-the-art performance on an existing language-conditioned folding benchmark. Given the lack of annotated bimanual folding data, we devise a procedure to automatically parse actions of a simulated dataset and tag them with aligned text instructions. BiFold attains the best performance on our dataset and can transfer to new instructions, garments, and environments.
2501.16466
On the Feasibility of Using LLMs to Execute Multistage Network Attacks
cs.CR cs.AI
LLMs have shown preliminary promise in some security tasks and CTF challenges. However, it is unclear whether LLMs are able to realize multistage network attacks, which involve executing a wide variety of actions across multiple hosts such as conducting reconnaissance, exploiting vulnerabilities to gain initial access, leveraging internal hosts to move laterally, and using multiple compromised hosts to exfiltrate data. We evaluate LLMs across 10 multistage networks and find that popular LLMs are unable to realize these attacks. To enable LLMs to realize these attacks, we introduce Incalmo, an LLM-agnostic high-level attack abstraction layer that sits between an LLM and the environment. Rather than LLMs issuing low-level command-line instructions, which can lead to incorrect implementations, Incalmo allows LLMs to specify high-level tasks (e.g., infect a host, scan a network), which are then carried out by Incalmo. Incalmo realizes these tasks by translating them into low-level primitives (e.g., commands to exploit tools). Incalmo also provides an environment state service and an attack graph service to provide structure to LLMs in selecting actions relevant to a multistage attack. Across 9 out of 10 realistic emulated networks (from 25 to 50 hosts), LLMs using Incalmo can successfully autonomously execute multistage attacks. We also conduct an ablation analysis to show the key role the high-level abstractions play. For instance, we find that both Incalmo's high-level tasks and services are crucial. Furthermore, even smaller-parameter LLMs with Incalmo can fully succeed in 5 of 10 environments, while larger-parameter LLMs without Incalmo do not fully succeed in any.
2501.16467
Cross-Domain Semantic Segmentation with Large Language Model-Assisted Descriptor Generation
cs.CV
Semantic segmentation plays a crucial role in enabling machines to understand and interpret visual scenes at a pixel level. While traditional segmentation methods have achieved remarkable success, their generalization to diverse scenes and unseen object categories remains limited. Recent advancements in large language models (LLMs) offer a promising avenue for bridging visual and textual modalities, providing a deeper understanding of semantic relationships. In this paper, we propose LangSeg, a novel LLM-guided semantic segmentation method that leverages context-sensitive, fine-grained subclass descriptors generated by LLMs. Our framework integrates these descriptors with a pre-trained Vision Transformer (ViT) to achieve superior segmentation performance without extensive model retraining. We evaluate LangSeg on two challenging datasets, ADE20K and COCO-Stuff, where it outperforms state-of-the-art models, achieving up to a 6.1% improvement in mean Intersection over Union (mIoU). Additionally, we conduct a comprehensive ablation study and human evaluation to validate the effectiveness of our method in real-world scenarios. The results demonstrate that LangSeg not only excels in semantic understanding and contextual alignment but also provides a flexible and efficient framework for language-guided segmentation tasks. This approach opens up new possibilities for interactive and domain-specific segmentation applications.
2501.16469
Object Detection for Medical Image Analysis: Insights from the RT-DETR Model
cs.CV cs.LG
Deep learning has emerged as a transformative approach for solving complex pattern recognition and object detection challenges. This paper focuses on the application of a novel detection framework based on the RT-DETR model for analyzing intricate image data, particularly in areas such as diabetic retinopathy detection. Diabetic retinopathy, a leading cause of vision loss globally, requires accurate and efficient image analysis to identify early-stage lesions. The proposed RT-DETR model, built on a Transformer-based architecture, excels at processing high-dimensional and complex visual data with enhanced robustness and accuracy. Comparative evaluations with models such as YOLOv5, YOLOv8, SSD, and DETR demonstrate that RT-DETR achieves superior performance across precision, recall, mAP50, and mAP50-95 metrics, particularly in detecting small-scale objects and densely packed targets. This study underscores the potential of Transformer-based models like RT-DETR for advancing object detection tasks, offering promising applications in medical imaging and beyond.
2501.16471
SIM: Surface-based fMRI Analysis for Inter-Subject Multimodal Decoding from Movie-Watching Experiments
cs.LG cs.AI eess.AS eess.IV q-bio.NC
Current AI frameworks for brain decoding and encoding, typically train and test models within the same datasets. This limits their utility for brain computer interfaces (BCI) or neurofeedback, for which it would be useful to pool experiences across individuals to better simulate stimuli not sampled during training. A key obstacle to model generalisation is the degree of variability of inter-subject cortical organisation, which makes it difficult to align or compare cortical signals across participants. In this paper we address this through the use of surface vision transformers, which build a generalisable model of cortical functional dynamics, through encoding the topography of cortical networks and their interactions as a moving image across a surface. This is then combined with tri-modal self-supervised contrastive (CLIP) alignment of audio, video, and fMRI modalities to enable the retrieval of visual and auditory stimuli from patterns of cortical activity (and vice-versa). We validate our approach on 7T task-fMRI data from 174 healthy participants engaged in the movie-watching experiment from the Human Connectome Project (HCP). Results show that it is possible to detect which movie clips an individual is watching purely from their brain activity, even for individuals and movies not seen during training. Further analysis of attention maps reveals that our model captures individual patterns of brain activity that reflect semantic and visual systems. This opens the door to future personalised simulations of brain function. Code & pre-trained models will be made available at https://github.com/metrics-lab/sim, processed data for training will be available upon request at https://gin.g-node.org/Sdahan30/sim.
2501.16473
Sensitivity Analysis of the Laser Power Control System to Measurement Noise in SLS 3D Printers
eess.SY cs.SY
Uniform temperature distribution in Selective Laser Sintering (SLS) is essential for producing durable 3D prints. Achieving uniformity requires a laser power control system that minimises deviation of the printing temperatures from the target temperature. Because the estimate of the actual process temperature is an input to the laser power control, uncertainty in the estimate of the actual temperature can lead to fluctuations in laser power that affect the thermal performance of the SLS. This article investigates the sensitivity of a laser power control system to temperature measurement uncertainty. This article evaluates the effectiveness of two methods for quantifying the effect of input uncertainty on a SLS laser power control system: a recent innovation in uncertainty-tracked architecture and traditional Monte Carlo simulation. We show that recent advances in computer architecture for arithmatic on probability distributions make it possible for the first time, to perform control system uncertainty analysis with latencies under 30 ms, while achieving the same level of uncertainty analysis as Monte Carlo methods with latencies that are two orders of magnitude slower.
2501.16476
Closed-Form Feedback-Free Learning with Forward Projection
cs.LG stat.ML
State-of-the-art methods for backpropagation-free learning employ local error feedback to direct iterative optimisation via gradient descent. In this study, we examine the more restrictive setting where retrograde communication from neuronal outputs is unavailable for pre-synaptic weight optimisation. To address this challenge, we propose Forward Projection (FP). This novel randomised closed-form training method requires only a single forward pass over the entire dataset for model fitting, without retrograde communication. Target values for pre-activation membrane potentials are generated layer-wise via nonlinear projections of pre-synaptic inputs and the labels. Local loss functions are optimised over pre-synaptic inputs using closed-form regression, without feedback from neuronal outputs or downstream layers. Interpretability is a key advantage of FP training; membrane potentials of hidden neurons in FP-trained networks encode information which is interpretable layer-wise as label predictions. We demonstrate the effectiveness of FP across four biomedical datasets. In few-shot learning tasks, FP yielded more generalisable models than those optimised via backpropagation. In large-sample tasks, FP-based models achieve generalisation comparable to gradient descent-based local learning methods while requiring only a single forward propagation step, achieving significant speed up for training. Interpretation functions defined on local neuronal activity in FP-based models successfully identified clinically salient features for diagnosis in two biomedical datasets. Forward Projection is a computationally efficient machine learning approach that yields interpretable neural network models without retrograde communication of neuronal activity during training.
2501.16480
Modular Framework for Uncertainty Prediction in Autonomous Vehicle Motion Forecasting within Complex Traffic Scenarios
cs.RO cs.LG eess.SP
We propose a modular modeling framework designed to enhance the capture and validation of uncertainty in autonomous vehicle (AV) trajectory prediction. Departing from traditional deterministic methods, our approach employs a flexible, end-to-end differentiable probabilistic encoder-decoder architecture. This modular design allows the encoder and decoder to be trained independently, enabling seamless adaptation to diverse traffic scenarios without retraining the entire system. Our key contributions include: (1) a probabilistic heatmap predictor that generates context-aware occupancy grids for dynamic forecasting, (2) a modular training approach that supports independent component training and flexible adaptation, and (3) a structured validation scheme leveraging uncertainty metrics to evaluate robustness under high-risk conditions. To highlight the benefits of our framework, we benchmark it against an end-to-end baseline, demonstrating faster convergence, improved stability, and flexibility. Experimental results validate these advantages, showcasing the capacity of the framework to efficiently handle complex scenarios while ensuring reliable predictions and robust uncertainty representation. This modular design offers significant practical utility and scalability for real-world autonomous driving applications.
2501.16481
Generating customized prompts for Zero-Shot Rare Event Medical Image Classification using LLM
cs.CV
Rare events, due to their infrequent occurrences, do not have much data, and hence deep learning techniques fail in estimating the distribution for such data. Open-vocabulary models represent an innovative approach to image classification. Unlike traditional models, these models classify images into any set of categories specified with natural language prompts during inference. These prompts usually comprise manually crafted templates (e.g., 'a photo of a {}') that are filled in with the names of each category. This paper introduces a simple yet effective method for generating highly accurate and contextually descriptive prompts containing discriminative characteristics. Rare event detection, especially in medicine, is more challenging due to low inter-class and high intra-class variability. To address these, we propose a novel approach that uses domain-specific expert knowledge on rare events to generate customized and contextually relevant prompts, which are then used by large language models for image classification. Our zero-shot, privacy-preserving method enhances rare event classification without additional training, outperforming state-of-the-art techniques.
2501.16485
Enhanced Position Estimation in Tactile Internet-Enabled Remote Robotic Surgery Using MOESP-Based Kalman Filter
cs.RO cs.SY eess.SY
Accurately estimating the position of a patient's side robotic arm in real time during remote surgery is a significant challenge, especially within Tactile Internet (TI) environments. This paper presents a new and efficient method for position estimation using a Kalman Filter (KF) combined with the Multivariable Output-Error State Space (MOESP) method for system identification. Unlike traditional approaches that require prior knowledge of the system's dynamics, this study uses the JIGSAW dataset, a comprehensive collection of robotic surgical data, along with input from the Master Tool Manipulator (MTM) to derive the state-space model directly. The MOESP method allows accurate modeling of the Patient Side Manipulator (PSM) dynamics without prior system models, improving the KF's performance under simulated network conditions, including delays, jitter, and packet loss. These conditions mimic real-world challenges in Tactile Internet applications. The findings demonstrate the KF's improved resilience and accuracy in state estimation, achieving over 95 percent accuracy despite network-induced uncertainties.
2501.16487
Network Risk Estimation: A Risk Estimation Paradigm for Cyber Networks
eess.SY cs.SY
Cyber networks are fundamental to many organization's infrastructure, and the size of cyber networks is increasing rapidly. Risk measurement of the entities/endpoints that make up the network via available knowledge about possible threats has been the primary tool in cyber network security. However, the dynamic behavior of the entities and the sparsity of risk-measurable points are limiting factors for risk measurement strategies, which results in poor network visibility considering the volatility of cyber networks. This work proposes a new probabilistic risk estimation approach to network security, NRE, which operates on top of existing risk measurements. The proposed method NRE extracts relationships among system components from the network connection data, models risk propagation based on the learned relationships and refines the estimates whenever risk measurements are provided. In this work, (i) the risk estimation scheme is proposed, (ii) an application of quantitative risk estimates is devised, (iii) descriptiveness of the risk estimates are compared to a pure risk measurement alternative and (iv) low computational complexity of the proposed method is illustrated capable of real-time deployment. The proposed method, NRE, is ultimately a quantitative data-driven risk assessment tool that can be used to add security aspects to existing network functions, such as routing, and it provides a robust description of the network state in the presence of threats, capable of running in real-time.
2501.16489
Nonparametric Sparse Online Learning of the Koopman Operator
stat.ML cs.LG cs.SY eess.SY
The Koopman operator provides a powerful framework for representing the dynamics of general nonlinear dynamical systems. Data-driven techniques to learn the Koopman operator typically assume that the chosen function space is closed under system dynamics. In this paper, we study the Koopman operator via its action on the reproducing kernel Hilbert space (RKHS), and explore the mis-specified scenario where the dynamics may escape the chosen function space. We relate the Koopman operator to the conditional mean embeddings (CME) operator and then present an operator stochastic approximation algorithm to learn the Koopman operator iteratively with control over the complexity of the representation. We provide both asymptotic and finite-time last-iterate guarantees of the online sparse learning algorithm with trajectory-based sampling with an analysis that is substantially more involved than that for finite-dimensional stochastic approximation. Numerical examples confirm the effectiveness of the proposed algorithm.
2501.16490
Towards Robust Stability Prediction in Smart Grids: GAN-based Approach under Data Constraints and Adversarial Challenges
cs.CR cs.AI cs.LG
Smart grids are critical for addressing the growing energy demand due to global population growth and urbanization. They enhance efficiency, reliability, and sustainability by integrating renewable energy. Ensuring their availability and safety requires advanced operational control and safety measures. Researchers employ AI and machine learning to assess grid stability, but challenges like the lack of datasets and cybersecurity threats, including adversarial attacks, persist. In particular, data scarcity is a key issue: obtaining grid instability instances is tough due to the need for significant expertise, resources, and time. However, they are essential to test novel research advancements and security mitigations. In this paper, we introduce a novel framework to detect instability in smart grids by employing only stable data. It relies on a Generative Adversarial Network (GAN) where the generator is trained to create instability data that are used along with stable data to train the discriminator. Moreover, we include a new adversarial training layer to improve robustness against adversarial attacks. Our solution, tested on a dataset composed of real-world stable and unstable samples, achieve accuracy up to 97.5\% in predicting grid stability and up to 98.9\% in detecting adversarial attacks. Moreover, we implemented our model in a single-board computer demonstrating efficient real-time decision-making with an average response time of less than 7ms. Our solution improves prediction accuracy and resilience while addressing data scarcity in smart grid management.
2501.16496
Open Problems in Mechanistic Interpretability
cs.LG
Mechanistic interpretability aims to understand the computational mechanisms underlying neural networks' capabilities in order to accomplish concrete scientific and engineering goals. Progress in this field thus promises to provide greater assurance over AI system behavior and shed light on exciting scientific questions about the nature of intelligence. Despite recent progress toward these goals, there are many open problems in the field that require solutions before many scientific and practical benefits can be realized: Our methods require both conceptual and practical improvements to reveal deeper insights; we must figure out how best to apply our methods in pursuit of specific goals; and the field must grapple with socio-technical challenges that influence and are influenced by our work. This forward-facing review discusses the current frontier of mechanistic interpretability and the open problems that the field may benefit from prioritizing.
2501.16497
Smoothed Embeddings for Robust Language Models
cs.LG cs.AI cs.CL cs.CR stat.ML
Improving the safety and reliability of large language models (LLMs) is a crucial aspect of realizing trustworthy AI systems. Although alignment methods aim to suppress harmful content generation, LLMs are often still vulnerable to jailbreaking attacks that employ adversarial inputs that subvert alignment and induce harmful outputs. We propose the Randomized Embedding Smoothing and Token Aggregation (RESTA) defense, which adds random noise to the embedding vectors and performs aggregation during the generation of each output token, with the aim of better preserving semantic information. Our experiments demonstrate that our approach achieves superior robustness versus utility tradeoffs compared to the baseline defenses.
2501.16504
Digital Twin Enabled Site Specific Channel Precoding: Over the Air CIR Inference
eess.SP cs.AI
This paper investigates the significance of designing a reliable, intelligent, and true physical environment-aware precoding scheme by leveraging an accurately designed channel twin model to obtain realistic channel state information (CSI) for cellular communication systems. Specifically, we propose a fine-tuned multi-step channel twin design process that can render CSI very close to the CSI of the actual environment. After generating a precise CSI, we execute precoding using the obtained CSI at the transmitter end. We demonstrate a two-step parameters' tuning approach to design channel twin by ray tracing (RT) emulation, then further fine-tuning of CSI by employing an artificial intelligence (AI) based algorithm can significantly reduce the gap between actual CSI and the fine-tuned digital twin (DT) rendered CSI. The simulation results show the effectiveness of the proposed novel approach in designing a true physical environment-aware channel twin model.
2501.16507
Characterizing Network Structure of Anti-Trans Actors on TikTok
cs.HC cs.AI cs.SI
The recent proliferation of short form video social media sites such as TikTok has been effectively utilized for increased visibility, communication, and community connection amongst trans/nonbinary creators online. However, these same platforms have also been exploited by right-wing actors targeting trans/nonbinary people, enabling such anti-trans actors to efficiently spread hate speech and propaganda. Given these divergent groups, what are the differences in network structure between anti-trans and pro-trans communities on TikTok, and to what extent do they amplify the effects of anti-trans content? In this paper, we collect a sample of TikTok videos containing pro and anti-trans content, and develop a taxonomy of trans related sentiment to enable the classification of content on TikTok, and ultimately analyze the reply network structures of pro-trans and anti-trans communities. In order to accomplish this, we worked with hired expert data annotators from the trans/nonbinary community in order to generate a sample of highly accurately labeled data. From this subset, we utilized a novel classification pipeline leveraging Retrieval-Augmented Generation (RAG) with annotated examples and taxonomy definitions to classify content into pro-trans, anti-trans, or neutral categories. We find that incorporating our taxonomy and its logics into our classification engine results in improved ability to differentiate trans related content, and that Results from network analysis indicate many interactions between posters of pro-trans and anti-trans content exist, further demonstrating targeting of trans individuals, and demonstrating the need for better content moderation tools
2501.16509
Reinforcement Learning for Quantum Circuit Design: Using Matrix Representations
quant-ph cs.AI
Quantum computing promises advantages over classical computing. The manufacturing of quantum hardware is in the infancy stage, called the Noisy Intermediate-Scale Quantum (NISQ) era. A major challenge is automated quantum circuit design that map a quantum circuit to gates in a universal gate set. In this paper, we present a generic MDP modeling and employ Q-learning and DQN algorithms for quantum circuit design. By leveraging the power of deep reinforcement learning, we aim to provide an automatic and scalable approach over traditional hand-crafted heuristic methods.
2501.16510
Decrypting the temperature field in flow boiling with latent diffusion models
physics.flu-dyn cs.AI
This paper presents an innovative method using Latent Diffusion Models (LDMs) to generate temperature fields from phase indicator maps. By leveraging the BubbleML dataset from numerical simulations, the LDM translates phase field data into corresponding temperature distributions through a two-stage training process involving a vector-quantized variational autoencoder (VQVAE) and a denoising autoencoder. The resulting model effectively reconstructs complex temperature fields at interfaces. Spectral analysis indicates a high degree of agreement with ground truth data in the low to mid wavenumber ranges, even though some inconsistencies are observed at higher wavenumbers, suggesting areas for further enhancement. This machine learning approach significantly reduces the computational burden of traditional simulations and improves the precision of experimental calibration methods. Future work will focus on refining the model's ability to represent small-scale turbulence and expanding its applicability to a broader range of boiling conditions.
2501.16513
Deception in LLMs: Self-Preservation and Autonomous Goals in Large Language Models
cs.CL
Recent advances in Large Language Models (LLMs) have incorporated planning and reasoning capabilities, enabling models to outline steps before execution and provide transparent reasoning paths. This enhancement has reduced errors in mathematical and logical tasks while improving accuracy. These developments have facilitated LLMs' use as agents that can interact with tools and adapt their responses based on new information. Our study examines DeepSeek R1, a model trained to output reasoning tokens similar to OpenAI's o1. Testing revealed concerning behaviors: the model exhibited deceptive tendencies and demonstrated self-preservation instincts, including attempts of self-replication, despite these traits not being explicitly programmed (or prompted). These findings raise concerns about LLMs potentially masking their true objectives behind a facade of alignment. When integrating such LLMs into robotic systems, the risks become tangible - a physically embodied AI exhibiting deceptive behaviors and self-preservation instincts could pursue its hidden objectives through real-world actions. This highlights the critical need for robust goal specification and safety frameworks before any physical implementation.
2501.16516
How well can LLMs Grade Essays in Arabic?
cs.CL cs.AI
This research assesses the effectiveness of state-of-the-art large language models (LLMs), including ChatGPT, Llama, Aya, Jais, and ACEGPT, in the task of Arabic automated essay scoring (AES) using the AR-AES dataset. It explores various evaluation methodologies, including zero-shot, few-shot in-context learning, and fine-tuning, and examines the influence of instruction-following capabilities through the inclusion of marking guidelines within the prompts. A mixed-language prompting strategy, integrating English prompts with Arabic content, was implemented to improve model comprehension and performance. Among the models tested, ACEGPT demonstrated the strongest performance across the dataset, achieving a Quadratic Weighted Kappa (QWK) of 0.67, but was outperformed by a smaller BERT-based model with a QWK of 0.88. The study identifies challenges faced by LLMs in processing Arabic, including tokenization complexities and higher computational demands. Performance variation across different courses underscores the need for adaptive models capable of handling diverse assessment formats and highlights the positive impact of effective prompt engineering on improving LLM outputs. To the best of our knowledge, this study is the first to empirically evaluate the performance of multiple generative Large Language Models (LLMs) on Arabic essays using authentic student data.
2501.16519
Optimizing Decentralized Online Learning for Supervised Regression and Classification Problems
cs.LG cs.DC cs.MA
Decentralized learning networks aim to synthesize a single network inference from a set of raw inferences provided by multiple participants. To determine the combined inference, these networks must adopt a mapping from historical participant performance to weights, and to appropriately incentivize contributions they must adopt a mapping from performance to fair rewards. Despite the increased prevalence of decentralized learning networks, there exists no systematic study that performs a calibration of the associated free parameters. Here we present an optimization framework for key parameters governing decentralized online learning in supervised regression and classification problems. These parameters include the slope of the mapping between historical performance and participant weight, the timeframe for performance evaluation, and the slope of the mapping between performance and rewards. These parameters are optimized using a suite of numerical experiments that mimic the design of the Allora Network, but have been extended to handle classification tasks in addition to regression tasks. This setup enables a comparative analysis of parameter tuning and network performance optimization (loss minimization) across both problem types. We demonstrate how the optimal performance-weight mapping, performance timeframe, and performance-reward mapping vary with network composition and problem type. Our findings provide valuable insights for the optimization of decentralized learning protocols, and we discuss how these results can be generalized to optimize any inference synthesis-based, decentralized AI network.
2501.16520
Safe Gradient Flow for Bilevel Optimization
math.OC cs.LG cs.SY eess.SY
Bilevel optimization is a key framework in hierarchical decision-making, where one problem is embedded within the constraints of another. In this work, we propose a control-theoretic approach to solving bilevel optimization problems. Our method consists of two components: a gradient flow mechanism to minimize the upper-level objective and a safety filter to enforce the constraints imposed by the lower-level problem. Together, these components form a safe gradient flow that solves the bilevel problem in a single loop. To improve scalability with respect to the lower-level problem's dimensions, we introduce a relaxed formulation and design a compact variant of the safe gradient flow. This variant minimizes the upper-level objective while ensuring the lower-level solution remains within a user-defined distance. Using Lyapunov analysis, we establish convergence guarantees for the dynamics, proving that they converge to a neighborhood of the optimal solution. Numerical experiments further validate the effectiveness of the proposed approaches. Our contributions provide both theoretical insights and practical tools for efficiently solving bilevel optimization problems.
2501.16524
Programming by Examples Meets Historical Linguistics: A Large Language Model Based Approach to Sound Law Induction
cs.CL
Historical linguists have long written "programs" that convert reconstructed words in an ancestor language into their attested descendants via ordered string rewrite functions (called sound laws) However, writing these programs is time-consuming, motivating the development of automated Sound Law Induction (SLI) which we formulate as Programming by Examples (PBE) with Large Language Models (LLMs) in this paper. While LLMs have been effective for code generation, recent work has shown that PBE is challenging but improvable by fine-tuning, especially with training data drawn from the same distribution as evaluation data. In this paper, we create a conceptual framework of what constitutes a "similar distribution" for SLI and propose four kinds of synthetic data generation methods with varying amounts of inductive bias to investigate what leads to the best performance. Based on the results we create a SOTA open-source model for SLI as PBE (+6% pass rate with a third of the parameters of the second-best LLM) and also highlight exciting future directions for PBE research.
2501.16525
Multi-Objective Deep-Learning-based Biomechanical Deformable Image Registration with MOREA
cs.CV cs.AI cs.NE
When choosing a deformable image registration (DIR) approach for images with large deformations and content mismatch, the realism of found transformations often needs to be traded off against the required runtime. DIR approaches using deep learning (DL) techniques have shown remarkable promise in instantly predicting a transformation. However, on difficult registration problems, the realism of these transformations can fall short. DIR approaches using biomechanical, finite element modeling (FEM) techniques can find more realistic transformations, but tend to require much longer runtimes. This work proposes the first hybrid approach to combine them, with the aim of getting the best of both worlds. This hybrid approach, called DL-MOREA, combines a recently introduced multi-objective DL-based DIR approach which leverages the VoxelMorph framework, called DL-MODIR, with MOREA, an evolutionary algorithm-based, multi-objective DIR approach in which a FEM-like biomechanical mesh transformation model is used. In our proposed hybrid approach, the DL results are used to smartly initialize MOREA, with the aim of more efficiently optimizing its mesh transformation model. We empirically compare DL-MOREA against its components, DL-MODIR and MOREA, on CT scan pairs capturing large bladder filling differences of 15 cervical cancer patients. While MOREA requires a median runtime of 45 minutes, DL-MOREA can already find high-quality transformations after 5 minutes. Compared to the DL-MODIR transformations, the transformations found by DL-MOREA exhibit far less folding and improve or preserve the bladder contour distance error.
2501.16533
A comparison of data filtering techniques for English-Polish LLM-based machine translation in the biomedical domain
cs.CL cs.LG
Large Language Models (LLMs) have become state-of-the-art in Machine Translation (MT), often trained on massive bilingual parallel corpora scraped from the web, that contain low-quality entries and redundant information, leading to significant computational challenges. Various data filtering methods exist to reduce dataset sizes, but their effectiveness largely varies based on specific language pairs and domains. This paper evaluates the impact of commonly used data filtering techniques, such as LASER, MUSE, and LaBSE, on English-Polish translation within the biomedical domain. By filtering the UFAL Medical Corpus, we created varying dataset sizes to fine-tune the mBART50 model, which was then evaluated using the SacreBLEU metric on the Khresmoi dataset, having the quality of translations assessed by bilingual speakers. Our results show that both LASER and MUSE can significantly reduce dataset sizes while maintaining or even enhancing performance. We recommend the use of LASER, as it consistently outperforms the other methods and provides the most fluent and natural-sounding translations.
2501.16534
Targeting Alignment: Extracting Safety Classifiers of Aligned LLMs
cs.CR cs.AI
Alignment in large language models (LLMs) is used to enforce guidelines such as safety. Yet, alignment fails in the face of jailbreak attacks that modify inputs to induce unsafe outputs. In this paper, we present and evaluate a method to assess the robustness of LLM alignment. We observe that alignment embeds a safety classifier in the target model that is responsible for deciding between refusal and compliance. We seek to extract an approximation of this classifier, called a surrogate classifier, from the LLM. We develop an algorithm for identifying candidate classifiers from subsets of the LLM model. We evaluate the degree to which the candidate classifiers approximate the model's embedded classifier in benign (F1 score) and adversarial (using surrogates in a white-box attack) settings. Our evaluation shows that the best candidates achieve accurate agreement (an F1 score above 80%) using as little as 20% of the model architecture. Further, we find attacks mounted on the surrogate models can be transferred with high accuracy. For example, a surrogate using only 50% of the Llama 2 model achieved an attack success rate (ASR) of 70%, a substantial improvement over attacking the LLM directly, where we only observed a 22% ASR. These results show that extracting surrogate classifiers is a viable (and highly effective) means for modeling (and therein addressing) the vulnerability of aligned models to jailbreaking attacks.