Author name cluster
Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.
EAAI Journal 2026 Journal Article
EAAI Journal 2026 Journal Article
NeurIPS Conference 2025 Conference Paper
Model editing has become a *de-facto* practice to address hallucinations and outdated knowledge of large language models (LLMs). However, existing methods are predominantly evaluated in isolation, i. e. , one edit at a time, failing to consider a critical scenario of compositional model editing, where multiple edits must be integrated and jointly utilized to answer real-world multifaceted questions. For instance, in medical domains, if one edit informs LLMs that COVID-19 causes "fever" and another that it causes "loss of taste", a qualified compositional editor should enable LLMs to answer the question "What are the symptoms of COVID-19? " with both "fever" and "loss of taste" (and potentially more). In this work, we define and systematically benchmark this compositional model editing (CME) task, identifying three key undesirable issues that existing methods struggle with: *knowledge loss*, *incorrect preceding* and *knowledge sinking*. To overcome these issues, we propose A$^3$E, a novel compositional editor that (1) ***a**daptively combines and **a**daptively regularizes* pre-trained foundation knowledge in LLMs in the stage of edit training and (2) ***a**daptively merges* multiple edits to better meet compositional needs in the stage of edit composing. Extensive experiments demonstrate that A$^3$E improves the composability by at least 22. 45\% without sacrificing the performance of non-compositional model editing.
NeurIPS Conference 2025 Conference Paper
Federated Learning (FL) enables collaborative model training across distributed clients without data sharing, but its high computational and communication demands strain resource-constrained devices. While existing methods use dynamic pruning to improve efficiency by periodically adjusting sparse model topologies while maintaining sparsity, these approaches suffer from issues such as greedy adjustments, unstable topologies, and communication inefficiency, resulting in less robust models and suboptimal performance under data heterogeneity and partial client availability. To address these challenges, we propose Fed erated R obust pruning via combinatorial T hompson S ampling (FedRTS), a novel framework designed to develop robust sparse models. FedRTS enhances robustness and performance through its Thompson Sampling-based Adjustment (TSAdj) mechanism, which uses probabilistic decisions informed by stable and farsighted information, instead of deterministic decisions reliant on unstable and myopic information in previous methods. Extensive experiments demonstrate that FedRTS achieves state-of-the-art performance in computer vision and natural language processing tasks while reducing communication costs, particularly excelling in scenarios with heterogeneous data distributions and partial client participation. Our codes are available at: https: //github. com/Little0o0/FedRTS.
NeurIPS Conference 2025 Conference Paper
Temporally consistent depth estimation from stereo video is critical for real-world applications such as augmented reality, where inconsistent depth estimation disrupts the immersion of users. Despite its importance, this task remains challenging due to the difficulty in modeling long-term temporal consistency in a computationally efficient manner. Previous methods attempt to address this by aggregating spatio-temporal information but face a fundamental trade-off: limited temporal modeling provides only modest gains, whereas capturing long-range dependencies significantly increases computational cost. To address this limitation, we introduce a memory buffer for modeling long-range spatio-temporal consistency while achieving efficient dynamic stereo matching. Inspired by the two-stage decision-making process in humans, we propose a Pick-and-Play Memory (PPM) construction module for dynamic Stereo matching, dubbed as PPMStereo. PPM consists of a pick process that identifies the most relevant frames and a play process that weights the selected frames adaptively for spatio-temporal aggregation. This two-stage collaborative process maintains a compact yet highly informative memory buffer while achieving temporally consistent information aggregation. Extensive experiments validate the effectiveness of PPMStereo, demonstrating state-of-the-art performance in both accuracy and temporal consistency. Codes are available at \textcolor{blue}{https: //github. com/cocowy1/PPMStereo}.
NeurIPS Conference 2025 Conference Paper
Safety validation, which assesses the safety of an autonomous system's motion planning decisions, is critical for the safe deployment of autonomous vehicles. Existing input validation techniques from other machine learning domains, such as image classification, face unique challenges in motion planning due to its contextual properties, including complex inputs and one-to-many mapping. Furthermore, current output validation methods in autonomous driving primarily focus on open-loop trajectory prediction, which is ill-suited for the closed-loop nature of motion planning. We introduce REDOUBT, the first systematic safety validation framework for autonomous vehicle motion planning that employs a duo mechanism, simultaneously inspecting input distributions and output uncertainty. REDOUBT identifies previously overlooked unsafe modes arising from the interplay of In-Distribution/Out-of-Distribution (OOD) scenarios and certain/uncertain planning decisions. We develop specialized solutions for both OOD detection via latent flow matching and decision uncertainty estimation via an energy-based approach. Our extensive experiments demonstrate that both modules outperform existing approaches, under both open-loop and closed-loop evaluation settings. Our codes are available at: https: //github. com/sgNicola/Redoubt.
JBHI Journal 2024 Journal Article
Psychophysiological computing can be utilized to analyze heterogeneous physiological signals with psychological behaviors in the Internet of Medical Things (IoMT). Since IoMT devices are generally limited by power, storage, and computing resources, it's very challenging to process the physiological signal securely and efficiently. In this work, we design a novel scheme named Heterogeneous Compression and Encryption Neural Network (HCEN), which aims to protect signal security and reduce the required resources in processing heterogeneous physiological signals. The proposed HCEN is designed as an integrated structure that introduces the adversarial properties of Generative Adversarial Networks (GAN) and the feature extraction functionality of Autoencoder (AE). Moreover, we conduct simulations to validate the performance of HCEN using the MIMIC-III waveform dataset. Electrocardiogram (ECG) and Photoplethysmography (PPG) signals are extracted in the simulation. The results reveal that the proposed HCEN can effectively encrypt floating-point signals. Meanwhile, the compression performance outperforms baseline compression methods.
JBHI Journal 2023 Journal Article
Federated learning (FL) is receiving much attention in the Healthcare Internet of Things (H-IoT) to support various instantaneous E-health services. Today, the deployment of FL suffers from several challenges, such as high training latency and data privacy leakage risks, especially for resource-constrained medical devices. In this article, we develop a three-layer FL architecture to decrease training latency by introducing split learning into FL. We formulate a long-term optimization problem to minimize the local model training latency while preserving the privacy of the original medical data in H-IoT. Specially, a Privacy-ware Model Partitioning Algorithm (PMPA) is proposed to solve the formulated problem based on the Lyapunov optimization theory. In PMPA, the local model is partitioned properly between a resource-constrained medical end device and an edge server, which meets privacy requirements and energy consumption constraints. The proposed PMPA is separated into two phases. In the first phase, a partition point set is obtained using Kullback-Leibler (KL) divergence to meet the privacy requirement. In the second phase, we employ the model partitioning function, derived through Lyapunov optimization, to select the partition point from the partition point set that that satisfies the energy consumption constraints. Simulation results show that compared with traditional FL, the proposed algorithm can significantly reduce the local training latency. Moreover, the proposed algorithm improves the efficiency of medical image classification while ensuring medical data security.
AAAI Conference 2019 Conference Paper
This paper proposes a novel neural architecture — Attentive Tensor Product Learning (ATPL) — to represent grammatical structures of natural language in deep learning models. ATPL exploits Tensor Product Representations (TPR), a structured neural-symbolic model developed in cognitive science, to integrate deep learning with explicit natural language structures and rules. The key ideas of ATPL are: 1) unsupervised learning of role-unbinding vectors of words via the TPR-based deep neural network; 2) the use of attention modules to compute TPR; and 3) the integration of TPR with typical deep learning architectures including long short-term memory and feedforward neural networks. The novelty of our approach lies in its ability to extract the grammatical structure of a sentence by using role-unbinding vectors, which are obtained in an unsupervised manner. Our ATPL approach is applied to 1) image captioning, 2) part of speech (POS) tagging, and 3) constituency parsing of a natural language sentence. The experimental results demonstrate the effectiveness of the proposed approach in all these three natural language processing tasks.
AAAI Conference 2019 Conference Paper
We propose a hierarchically structured reinforcement learning approach to address the challenges of planning for generating coherent multi-sentence stories for the visual storytelling task. Within our framework, the task of generating a story given a sequence of images is divided across a two-level hierarchical decoder. The high-level decoder constructs a plan by generating a semantic concept (i. e. , topic) for each image in sequence. The low-level decoder generates a sentence for each image using a semantic compositional network, which effectively grounds the sentence generation conditioned on the topic. The two decoders are jointly trained end-to-end using reinforcement learning. We evaluate our model on the visual storytelling (VIST) dataset. Empirical results from both automatic and human evaluations demonstrate that the proposed hierarchically structured reinforced training achieves significantly better performance compared to a strong flat deep reinforcement learning baseline.
NeurIPS Conference 2018 Conference Paper
We study in this paper the problems of both image captioning and text-to-image generation, and present a novel turbo learning approach to jointly training an image-to-text generator (a. k. a. CaptionBot) and a text-to-image generator (a. k. a. DrawingBot). The key idea behind the joint training is that image-to-text generation and text-to-image generation as dual problems can form a closed loop to provide informative feedback to each other. Based on such feedback, we introduce a new loss metric by comparing the original input with the output produced by the closed loop. In addition to the old loss metrics used in CaptionBot and DrawingBot, this extra loss metric makes the jointly trained CaptionBot and DrawingBot better than the separately trained CaptionBot and DrawingBot. Furthermore, the turbo-learning approach enables semi-supervised learning since the closed loop can provide peudo-labels for unlabeled samples. Experimental results on the COCO dataset demonstrate that the proposed turbo learning can significantly improve the performance of both CaptionBot and DrawingBot by a large margin.
JMLR Journal 2013 Journal Article
Recently, researchers have proposed penalized maximum likelihood to identify network topology underlying a dynamical system modeled by multivariate time series. The time series of interest are assumed to be stationary, but this restriction is never taken into consideration by existing estimation methods. Moreover, practical problems of interest may have ultra-high dimensionality and obvious node collinearity. In addition, none of the available algorithms provides a probabilistic measure of the uncertainty for the obtained network topology which is informative in reliable network identification. The main purpose of this paper is to tackle these challenging issues. We propose the $\mathbf{S}^2$ learning framework, which stands for stationary- sparse network learning. We propose a novel algorithm referred to as the Berhu iterative sparsity pursuit with stationarity (BISPS), where the Berhu regularization can improve the Lasso in detection and estimation. The algorithm is extremely easy to implement, efficient in computation and has a theoretical guarantee to converge to a global optimum. We also incorporate a screening technique into BISPS to tackle ultra- high dimensional problems and enhance computational efficiency. Furthermore, a stationary bootstrap technique is applied to provide connection occurring frequency for reliable topology learning. Experiments show that our method can achieve stationary and sparse causality network learning and is scalable for high-dimensional problems. [abs] [ pdf ][ bib ] © JMLR 2013. ( edit, beta )