AI conference papers

AI conference paper library

Search papers from 30+ top conferences, including NeurIPS, ICLR, ICML, CVPR, ACL and AAAI. Filter by keyword, then open any paper for its abstract, authors and links.

arXiv

NORi: An ML-Augmented Ocean Boundary Layer Parameterization

Xin Kai Lee, Ali Ramadhan, Andre Souza, Gregory LeClaire Wagner, Simone Silvestri, John Marshall, Raffaele Ferrari

NORi is a machine learning (ML) parameterization of ocean boundary layer turbulence that is physics-based and augmented with neural networks. NORi stands for neural ordinary differential equations (NODEs) Richardson number (Ri) closure. The physical parameterization is controlled by a Richardson number-dependent diffusivity and viscosity. The neural ODEs are trained to capture the entrainment through the base of the boundary layer, which cannot be represented with a local diffusive closure. The parameterization is trained using large-eddy simulations in an a posteriori fashion, where parameters are calibrated with a loss function that explicitly depends on the actual time-integrated variables of interest rather than the instantaneous subgrid fluxes, which are inherently noisy. NORi conserves tracers by design, uses realistic nonlinear thermodynamics, and demonstrates excellent prediction and generalization capabilities in capturing entrainment dynamics under different convective strengths, background stratifications, rotation, and wind forcings. NORi is shown to simulate the seasonal evolution of the boundary layer at Ocean Weather Station Papa with similar performance to the state-of-the-art two-equation k-epsilon closure. When implemented in a double-gyre simulation, it is numerically stable for at least 100 years, despite only being trained on two-day horizons, and can be run with time steps as long as one hour. Combining highly expressive neural networks with a physically grounded base closure proves to be a robust paradigm for designing parameterizations for climate models: data required and training cost are drastically reduced, inference performance can be directly optimized as a primary objective, and numerical stability is implicitly promoted through training.

arXiv

Beyond Noise: A Hypothesis Testing Approach to Robust Feature Selection

Mousam Sinha, Tirtha Sarathi Ghosh, Koushik Biswas, Ridam Pal

Feature selection remains difficult in modern high-dimensional settings, and established methods such as Boruta and Recursive Feature Elimination are either computationally costly or lack a statistically justified stopping criterion for their importance scores. A common heuristic adds random noise features and retains any predictor ranking above the strongest one, but this rule is purely ad hoc. We introduce a method that keeps the noise-augmentation idea while grounding it in theory: each feature's importance is tested against the maximum noise importance using a non-parametric bootstrap hypothesis test, with statistical derivations supporting the algorithm's design. On controlled simulations, the method recovers true signal more consistently than Boruta and Knockoff-based procedures; on diverse real-world datasets, it outperforms Boruta, RFE, and Extra Trees. The result is a robust, principled selector that yields reliable inference, improved prediction, and efficient computation.

arXiv

Near-Optimal Reinforcement Learning for Constrained Recurrence Objectives

Dominik Wagner, Leon Witzman, Luke Ong

Recurrence objectives, where a target region must be visited infinitely often, are a fundamental class of specifications for Markov decision processes (MDPs) and form the core of $ω$-regular and linear temporal logic (LTL) objectives. We study constrained recurrence objectives, a natural extension of recurrence objectives with probabilistic constraints capable of modelling safety or fairness requirements. We first study the structure of optimal policies, showing that constrained recurrence requires different policy classes than those sufficient for other constrained MDP formalisms. In particular, we show that every feasible instance admits an optimal mixture of two stochastic stationary policies, as well as an optimal mixture of two deterministic stationary policies over a one-bit augmented MDP. We then study the generative-model reinforcement learning setting and propose an algorithm that first identifies the maximal end-component decomposition of the MDP, then reduces constrained recurrence to a constrained average reward problem for a collapsed MDP. Moreover, we establish a $\tilde{\mathcal{O}}(1/p+B/\varepsilon^2)$ sample complexity guarantee per state-action pair, where $p$ bounds certain non-zero transition probabilities and $B$ bounds transient time. Finally, we prove a nearly matching lower bound, showing that $1/p$ dependence, unlike the average-reward setting, is unavoidable.

arXiv

LAYA: Layer-wise Attention Aggregation for Interpretable Depth-Aware Neural Networks

Gennaro Vessio

Deep neural networks typically rely on the representation produced by their final hidden layer to make predictions, implicitly assuming that this single vector fully captures the semantics encoded across all preceding transformations. However, intermediate layers contain rich and complementary information---ranging from low-level patterns to high-level abstractions---that is often discarded when the decision head depends solely on the last representation. This paper revisits the role of the output layer and introduces LAYA (Layer-wise Attention Aggregator), a novel output head that dynamically aggregates internal representations through attention. Instead of projecting only the deepest embedding, LAYA learns input-conditioned attention weights over layer-wise features, yielding an interpretable and architecture-agnostic mechanism for synthesizing predictions. Beyond improving feature aggregation, the learned attention coefficients provide intrinsic layer-attribution scores that explicitly quantify the contribution of each representation to the final decision, without requiring external post-hoc explanation methods. Experiments on image classification datasets show that LAYA achieves competitive predictive performance while producing meaningful depth-aware explanations. Furthermore, quantitative and qualitative interpretability analyses demonstrate that LAYA's attention scores closely reflect the actual contribution of individual layers, revealing structured, task-dependent patterns of depth utilization while providing intuitive explanations of how different abstraction levels contribute to each prediction.

arXiv

Regularized Schrödinger Bridge via Distortion-Perception Perturbation for High-Fidelity Speech Enhancement

Qing Yao, Lijian Gao, Qirong Mao, Ming Dong

Speech enhancement (SE) requires high-fidelity reconstruction of clean speech that preserves linguistic and paralinguistic cues while maintaining high perceptual quality. Recently, Schrödinger Bridge (SB), a family of diffusion-based generative models, has advanced SE by bridging degraded and clean speech distributions in a principled formulation, enabling higher-quality reconstructions with fewer sampling steps. However, diffusion-based SE methods still face two challenges: (1) the fidelity-realism tradeoff, where they often prioritize perceptual realism encouraged by the learned speech prior, at the expense of fidelity; and (2) the exposure bias issue, where iterative multi-step sampling causes early-step prediction errors to accumulate along the sampling trajectory and degrade enhanced speech quality. In this paper, we analyze standard SB training and show that it induces a systematic prediction drift, which biases the multi-step trajectory and amplifies error accumulation. To address this, we propose Regularized Schrödinger Bridge (RSB) for high-fidelity SE, a generative approach that reconciles fidelity and realism while mitigating exposure bias. RSB regularizes training with a Distortion-Perception Perturbation that constructs time-varying targets by interpolating between clean speech and posterior-mean estimates, and trains the network on perturbed intermediate states to correct toward the ground truth progressively. By simulating inference-time prediction errors, this perturbation mitigates the training-inference mismatch and thereby alleviates exposure bias. It also injects posterior-mean estimates as fidelity-preserving guidance, thereby improving reconstruction fidelity.

arXiv

Eigenvalues as a Metric for Memory Dynamics in Sequence Models

Rahel Rickenbach, Jelena Trisovic, Alexandre Didier, Jerome Sieber, Melanie N. Zeilinger

While softmax attention drives state-of-the-art performance in sequence modeling, its quadratic complexity motivates linear alternatives such as state space models (SSMs). Structural differences between the two model classes, however, hinder direct comparisons of their memory dynamics, creating the need for a common metric to analyze, interpret, and improve their information processing capabilities. Inspired by recent advances in SSM performance driven by eigenvalue-guided insights, we leverage the dynamical systems framework to bring attention models into a unified analytical framework with SSMs. This allows us to perform a structured analysis, which investigates the applicability of an eigenvalue-spectrum memory dynamics metric to attention models. To this end, we first conduct an extensive empirical study across diverse attention-based models and SSMs on a range of benchmarks. We show that, for both model classes, eigenvalues influence key aspects of memory and long-range dependency modeling, revealing spectral signatures that align with task requirements. Building on these findings, we show how spectral signatures can motivate architectural modifications, how they can be guided through the training process, and how they can provide information about feature importance. The results thereby enable and emphasize the role of eigenvalue analysis as a principled metric for interpreting, explaining, and ultimately improving the capabilities of sequence models.

arXiv

Surrogate Modeling for the Design of Optimal Lattice Structures using Tensor Completion

Shaan Pakala, Aldair E. Gongora, Brian Giera, Evangelos E. Papalexakis

When designing new materials, it is often necessary to design a material with specific desired properties. Unfortunately, as new design variables are added, the search space grows exponentially, which makes synthesizing and validating the properties of each material very impractical and time-consuming. In this work, we focus on the design of optimal lattice structures with regard to mechanical performance. Computational approaches, including the use of machine learning (ML) methods, have shown improved success in accelerating materials design. However, these ML methods are still lacking in scenarios when training data (i.e. experimentally validated materials) come from a non-uniformly random sampling across the design space. For example, an experimentalist might synthesize and validate certain materials more frequently because of convenience. For this reason, we suggest the use of tensor completion as a surrogate model to accelerate the design of materials in these atypical supervised learning scenarios. In our experiments, we show that tensor completion is superior to classic ML methods such as Gaussian Process and XGBoost with biased sampling of the search space, with around 5\% increased $R^2$. Furthermore, tensor completion still gives comparable performance with a uniformly random sampling of the entire search space.

arXiv

Breaking the Exploration Bottleneck: Rubric-Scaffolded Reinforcement Learning for General LLM Reasoning

Yang Zhou, Sunzhu Li, Shunyu Liu, Wenkai Fang, Kongcheng Zhang, Jiale Zhao, Jingwen Yang, Yihe Zhou, Jianwei Lv, Tongya Zheng, Hengtong Lu, Wei Chen, Yan Xie, Mingli Song

Recent advances in Large Language Models (LLMs) have underscored the potential of Reinforcement Learning (RL) to facilitate the emergence of reasoning capabilities. Despite the encouraging results, a fundamental dilemma persists as RL improvement relies on learning from high-quality samples, yet the exploration for such samples remains bounded by the inherent limitations of LLMs. This, in effect, creates an undesirable cycle in which what cannot be explored cannot be learned. In this work, we propose Rubric-Scaffolded Reinforcement Learning (RuscaRL), a novel instructional scaffolding framework designed to break the exploration bottleneck for open-ended reasoning. Specifically, RuscaRL introduces checklist-style rubrics as (1) explicit scaffolding for exploration during rollout generation, where different rubrics are provided as external guidance within task instructions to steer diverse high-quality responses. This guidance is gradually decayed over time, encouraging the model to internalize the underlying reasoning patterns; (2) verifiable rewards for exploitation during model training, where we can obtain robust LLM-as-a-Judge scores using rubrics as references, enabling effective RL on open-ended reasoning tasks. Extensive experiments demonstrate the superiority of the proposed RuscaRL across various benchmarks, effectively expanding reasoning boundaries under the Best-of-N evaluation. Our code is available at https://github.com/IANNXANG/RuscaRL.

arXiv

Barron Space Representations for Elliptic PDEs with Homogeneous Boundary Conditions

Ziang Chen, Liqiang Huang

We study the complexity of approximating high-dimensional second-order elliptic PDEs with homogeneous boundary conditions on the unit hypercube using Barron spaces. Under suitable Barron assumptions on the coefficients and forcing term, we prove that the solutions can be approximated to any prescribed accuracy \(\varepsilon>0\) by two-layer neural networks whose widths and relevant parameters are bounded by \(\mathcal{O}\bigl(d^{C|\log\varepsilon|}\bigr)\). Consequently, we identify a class of elliptic PDEs that can be approximated by shallow neural networks without suffering from the curse of dimensionality.

arXiv

Quality-Diversity Red-Teaming: Automated Generation of High-Quality and Diverse Attackers for Large Language Models

Ren-Jian Wang, Ke Xue, Zeyu Qin, Ziniu Li, Sheng Tang, Hao-Tian Li, Shengcai Liu, Zhi Yu, Yuanpeng Tan, Chao Qian

Ensuring the safety and robustness of large language models (LLMs) is a fundamental challenge and a critical prerequisite for the responsible deployment of artificial intelligence. Red-teaming, a systematic framework to identify adversarial prompts that elicit harmful responses from target LLMs, has emerged as a crucial safety evaluation paradigm. Within this framework, the diversity of adversarial prompts is critical for a comprehensive safety assessment. However, previous red-teaming approaches often pursue diversity through simplistic metrics such as word frequency or sentence embedding similarity, which may not capture meaningful variation in attack strategies. In addition, the common practice of training a single attacker model restricts coverage across all potential attack styles and risk categories. This paper introduces Quality-Diversity Red-Teaming (QDRT), a new framework designed to address these limitations. QDRT achieves goal-driven diversity through behavior-conditioned training and implements a behavioral replay buffer in an open-ended manner. In addition, it trains multiple specialized attackers capable of generating high-quality attacks across diverse styles and risk categories. Our empirical evaluation demonstrates that QDRT generates attacks that are both more diverse and more effective against a wide range of target LLMs, including open-source models \texttt{GPT-2}, \texttt{Llama-3}, \texttt{Gemma-2}, \texttt{Qwen2.5}, and commercial models \texttt{GPT-4.1} and \texttt{GPT-5-Chat}. This work advances the field of LLM safety by providing a systematic and effective approach to automated red-teaming, ultimately supporting the responsible deployment of LLMs. Our code is available at https://github.com/lamda-bbo/QDRT.

arXiv

Credit Assignment and Efficient Exploration based on Influence Scope in Multi-agent Reinforcement Learning

Shuai Han, Mehdi Dastani, Shihan Wang

Training cooperative agents in sparse-reward scenarios poses significant challenges for multi-agent reinforcement learning (MARL). Without clear feedback on actions at each step in sparse-reward setting, previous methods struggle with precise credit assignment among agents and effective exploration. In this paper, we introduce a novel method to deal with both credit assignment and exploration problems in reward-sparse domains. Accordingly, we propose an algorithm that calculates the Influence Scope of Agents (ISA) on states by taking specific value of the dimensions/attributes of states that can be influenced by individual agents. The mutual dependence between agents' actions and state attributes are then used to calculate the credit assignment and to delimit the exploration space for each individual agent. We then evaluate ISA in a variety of sparse-reward multi-agent scenarios. The results show that our method significantly outperforms the state-of-art baselines.

arXiv

Fairness in Augmented Graph Learning: A Survey

Renqiang Luo, Huafei Huang, Ziqi Xu, Xikun Zhang, Enyan Dai, Bo Yang, Feng Xia

Graph learning has evolved into Augmented Graph Learning (AGL) by integrating specialized machine learning (ML) techniques. Examples include federated learning, graph transformers, and graph condensation. While enhancing model utility, AGL introduces unique intersectional fairness challenges that traditional GNN debiasing frameworks, which primarily focus on message-passing regulations, fail to address. This paper provides a systematic investigation into this emerging field, termed FairGX. We first delineate the shift from conventional fairness-aware graph learning to the FairGX paradigm, identifying novel bias sources inherent in ML augmentations, such as dual-side disparities in federated aggregation and attention-head skewness. A structured taxonomy is established to categorize existing literature based on their technical integration and fairness objectives. Furthermore, we analyze the impact of diverse ML paradigms on algorithmic equity, emphasizing the unique challenges in human-centered applications and the absence of a unified framework. We conclude by identifying five critical future directions, including novel metrics for AGL, fairness-privacy synergy, and Fairness-aware LLM4Graph/Graph4LLM. This survey serves as a foundational roadmap for developing robust and equitable graph systems in complex ML environments.

arXiv

Polyatomic Complexes: A topologically-informed learning representation for atomistic systems

Rahul Khorana, Marcus Noack, Jin Qian

A representation of a molecule or material should be invariant to the symmetries of physics, unique, continuous, efficient and general. These properties, however, are hard to satisfy at once: a descriptor invariant under the full orthogonal group $O(3)$ gives a molecule and its mirror image the same value, and so cannot distinguish enantiomers whose properties differ. Pozdnyakov showed this follows from the invariance itself, not from a lack of parameters. We show the criteria can be met at once if the geometric map is graded by the sign character of $O(3)$ and pooled multisymmetrically. We construct such a map $Φ$: its even block factors through the Gram matrix and is provably chirality-blind, while its parity-odd block of signed triple products separates enantiomers on an open dense full-measure set of interacting configurations. Two standard obstructions to uniqueness, fixed output length and componentwise pooling, are artifacts of the pooling rule, removed by multisymmetric power sums of order at most $N$. We establish uniqueness for a complete descriptor $Φ^\star$ built from the distance matrix, signed volumes and atom types, injective up to $SE(3)\times S_N$ on all configurations. $Φ^\star$ is non-constructive, however; the implemented map is the bounded-cutoff $Φ$, generically injective, pooling at order $2$, running in $O(N^2)$, or $O(N)$ with neighbor lists. The algebraic core is machine-checked in Lean 4. Because the underlying object is a cell complex, it also yields invariant, stable topological features $Ψ$ (persistent homology and a Hodge-Laplacian spectrum) encoding global ring and cage structure invisible to bounded-cutoff descriptors. The pair $(Φ,Ψ)$ feeds a compact parity-graded equivariant transformer.

arXiv

Simplified Quadratic Gradient: A Unified Framework Bridging Gradient Descent and Newton-Type Methods by Synthesizing Hessians and Gradients

John Chiang

Accelerating the convergence of second-order optimization, particularly Newton-type methods, remains a pivotal challenge in algorithmic research. In this paper, we extend previous work on the \textbf{Quadratic Gradient (QG)} and rigorously validate its applicability to general convex numerical optimization problems. We introduce a novel variant of the Quadratic Gradient that departs from the conventional fixed Hessian Newton framework. We present a new way to build a new version of the quadratic gradient. This new quadratic gradient doesn't satisfy the convergence conditions of the fixed Hessian Newton's method. However, experimental results show that it sometimes has a better performance than the original one in convergence rate. While this variant relaxes certain classical convergence constraints, it maintains a positive-definite Hessian proxy and demonstrates comparable, or in some cases superior, empirical performance in convergence rates. Furthermore, we demonstrate that both the original and the proposed QG variants can be effectively applied to non-convex optimization landscapes. A key motivation of our work is the limitation of traditional scalar learning rates. We argue that a diagonal matrix can more effectively accelerate gradient elements at heterogeneous rates. Our findings establish the Quadratic Gradient as a versatile and potent framework for modern optimization. Furthermore, we integrate Hutchinson's Estimator to estimate the Hessian diagonal efficiently via Hessian-vector products. Notably, we demonstrate that the proposed Quadratic Gradient variant is highly effective for Deep Learning architectures, providing a robust second-order alternative to standard adaptive optimizers.

arXiv

Query Timing Produces Opposite Positional Biases Between LLMs and Humans

Jasin Cekinmez, Addison J. Wu, Thomas L. Griffiths

Positional biases such as recency and primacy effects have been documented in large language models (LLMs), yet the underlying mechanism by which these models make their evaluations remains poorly understood. Both primacy and recency biases have been observed in human judgments in response to evidence, but recent work suggest that \emph{when} the listener updates their beliefs -- during the presentation of evidence or only at the end -- influences the presence of such effects. We investigate whether a similar phenomenon holds for LLMs, finding divergence from human behavior. These biases are more exacerbated in newer models compared to their predecessors.

arXiv

RSRA: Training-Free Probing of Representation Sensitivity for Efficient LoRA Rank Allocation

Jiaqi Liu, Haidong Kang, Qihui Zhao, Guo Yu, Jingchao Wang

Parameter-efficient fine-tuning enables large language models to adapt to downstream tasks with substantially lower computational and storage cost, and Low-Rank Adaptation (LoRA) is among its most widely used techniques. However, vanilla LoRA assigns a uniform rank to all adapted modules, while existing adaptive methods either incur additional optimization overhead or rely on static weights and local gradients that do not capture task-conditioned representation changes. We propose RSRA, a training-free rank allocator that estimates where adaptation capacity is most needed through forward-only representation sensitivity probing on a small calibration set. Specifically, RSRA uses Spectral Effective Rank to allocate capacity across layers, measures module-wise hidden-state displacement under standardized virtual low-rank updates with the Frechet Distance, and combines both signals through hierarchical normalization to produce a task-aware rank configuration before fine-tuning. Across commonsense reasoning and natural language understanding benchmarks with Qwen3-4B and Mistral-7B, RSRA achieves the highest average performance in all three reported model-benchmark settings and a 1.48x-1.93x speedup in allocation time over the fastest competing pre-allocation method. When integrated with DoRA, LoRA-FA, and PiSSA, RSRA improves 15 of the 18 evaluated combinations and increases the average performance of all three PEFT methods.

arXiv

NuclearDiffusion: Text-to-Image Foundation Models for Learning Nuclear Energy Concepts

Mohammed I. Radaideh, Jeremy Moon, Andre Gala-Garza, Emma Son, Yug Shah, Majdi I. Radaideh

Generative artificial intelligence (AI) has transformed text-to-image synthesis, yet its ability to represent specialized engineering domains remains largely unexplored. As an exmaple in nuclear engineering, general-purpose foundation models frequently generate physically incorrect or conceptually inconsistent images because they lack domain-specific knowledge. This work presents one of the first systematic studies of domain adaptation for nuclear text-to-image generation through fine-tuning of open-source diffusion models. We curate a dataset of 1,000 captioned nuclear energy images spanning reactors, fuel cycles, radiation, and related concepts, and use it to fine-tune three state-of-the-art open-source models: Stable Diffusion XL (SDXL), SD-v3.5-Medium, and the flow-matching Flux.1 model. Their performance is evaluated using both quantitative image-similarity metrics and qualitative expert assessment against the corresponding zero-shot models. Fine-tuning substantially improves the fidelity of SDXL, provides only limited gains for SD-v3.5-Medium, and yields no measurable improvement for Flux.1, demonstrating that adaptation effectiveness depends strongly on the underlying generative architecture rather than model scale alone. We further compare the fine-tuned models against three leading commercial systems--GPT-Image-2, Gemini-3.1-Flash-Image, and Midjourney. Although GPT-Image-2 and Gemini generate convincing images for broad nuclear concepts, they frequently fail on specialized engineering prompts, where the fine-tuned open-source models produce more accurate and technically consistent outputs. These results establish domain-specific fine-tuning as a practical pathway for developing trustworthy generative AI tools for domain-specific applications.

arXiv

Decoding Children's Gait Behavior

Yifan Shen, Boyi Li, Meihuan Huang, Yuanzhe Liu, Xu Cao, Jinyang Jin, Zhengyuan Li, Anglin Liu, Junho Kim, JingYuan Zhu, Lan Fangzhou, Jianguo Cao, Jintai Chen, Ismini Lourentzou, James Matthew Rehg

We introduce a new problem domain for human action recognition: the fine-grained analysis of children's gait behaviors from standard RGB video. We specifically target the ambulatory patterns of children aged 3-17 years. Such behaviors arise naturally in the diagnosis and treatment of several critical developmental and neuromuscular disorders, such as cerebral palsy and hemiplegia. Despite their clinical value, current 3D sensor-based gait analysis systems are expensive, intrusive, and often impractical for young subjects. To address this, we introduce a new dataset comprising over 1,100 high-frame-rate (60 FPS) video sequences from 110 subjects, accompanied by synchronized, anonymized pose sequences. In each session, the child performs a 5-second "walk-around" task, capturing the gait cycle from multiple viewpoints. Crucially, we demonstrate that current state-of-the-art approaches, including gait foundation models and Multimodal Large Language Models (MLLMs), fail to effectively resolve these clinical nuances. We identify the key technical challenges in analyzing these erratic and subtle motor patterns and describe a unified end-to-end framework for decoding fundamental components of pediatric gait. Through comprehensive experimental results, we demonstrate the potential of this dataset to drive novel research questions and establish a rigorous baseline for automated child gait assessment.

arXiv

MBO Scheme for Local Chan--Vese Segmentation

Kevin Bui, Adina Ciomaga

Robust to intensity inhomogeneity, the local Chan--Vese (LCV) model extends the classical Chan--Vese (CV) image segmentation method by incorporating local statistical information around each pixel. Originally, the LCV model was solved using a finite difference scheme, following the approach used for the CV model. As an alternative to the finite difference scheme, a more efficient algorithm based on the Merriman-Bence-Osher (MBO) scheme was later developed for the CV model. In this paper, we derive a similar MBO-based algorithm to solve the LCV model and propose an efficient implementation. The algorithm is developed for both two-phase and multiphase segmentation, and an extension to color images is also discussed. To demonstrate the effectiveness of the proposed approach, we apply it to a variety of grayscale and color images, including medical and microscopy images.

arXiv

Similarity Weighted Aggregation with Global Differential Privacy for Federated Brain Lesion Segmentation

Muhammad Irfan Khan, Eero Lehtonen, Joni Obradovic, Elina Kontio, Esa Alhoniemi, Suleiman A. Khan, Mojtaba Jafaritadi

Federated Learning (FL) enables collaborative training of machine learning models across multiple institutions without sharing sensitive data, making it particularly suitable for medical imaging applications. However, heterogeneous data distributions across institutions and potential information leakage through model updates remain important challenges. In this work, we propose DP-SimAgg, a privacy-preserving federated learning framework that integrates similarity-weighted aggregation with a server-side differential privacy mechanism. The proposed method applies L2 clipping to bound collaborator updates, computes similarity-based aggregation weights to mitigate the effects of non-IID data distributions, and injects calibrated Gaussian noise at the central server, providing per-round privacy guarantees under the assumed sensitivity bound. The framework is implemented using Intel's OpenFL platform and evaluated on the FeTS 2022 dataset consisting of 1251 multi-modal MRI scans for brain tumor segmentation. Experimental results demonstrate that DP-SimAgg maintains competitive segmentation performance while providing privacy protection. Under a strict per-round privacy budget (epsilon = 1, cumulative epsilon_total = 20 over 20 rounds), the method achieves Dice scores of 0.6357, 0.5305, and 0.5274 for the enhancing tumor (ET), tumor core (TC), and whole tumor (WT) regions, respectively. With a more relaxed per-round budget (epsilon = 10, cumulative epsilon_total = 200), performance approaches that of the non-private baseline while incorporating a central Gaussian mechanism with per-round (epsilon, delta)-DP accounting under the assumed sensitivity bound. These results highlight the potential of DP-SimAgg for enabling privacy-preserving collaborative learning in medical imaging applications.

Keep browsing

More papers from top AI conferences

30+ top conferences, by conference, year, and topic.