arXiv subjects

Jun Hong

Publications and source records attributed to Jun Hong.

14 recordsLinked to original sources

arXiv2026

AdaHAT: Adaptive Hard Attention to the Task in Task-Incremental Learning

Catastrophic forgetting is a major problem in task-incremental learning, where neural networks tend to overwrite previously learned knowledge when trained on new tasks. A number of architecture-based approaches have been proposed to address this problem. However, the architecture-based approaches suffer from another problem related to network capacity when the networks learn long task sequences: As a network is trained on an increasing number of new tasks in a long task sequence, a growing proportion of active parameters becomes static to prevent forgetting of previously learned knowledge. In this paper, we propose Adaptive Hard Attention to the Task (AdaHAT) with an adaptive attention mechanism which allows adaptive updates to static parameters by taking into account the information about previous tasks on both the importance of these parameters to previous tasks and the current network capacity. Based on this idea, we develop a new neural network architecture incorporating our proposed AdaHAT mechanism. AdaHAT extends an existing architecture-based approach, Hard Attention to the Task (HAT), to better support task-incremental learning over long task sequences. We conduct experiments on a number of datasets and compare AdaHAT with task-incremental learning baselines including HAT. Our experimental results show that AdaHAT achieves better average performance across tasks than these baselines, especially on long task sequences, demonstrating the benefits from balancing the trade-off between stability and plasticity of a network when learning such sequences of tasks, alleviating the network capacity problem. Our code is available at pengxiang-wang.com/projects/continual-learning-arena.

cs.LG↗

arXiv2025

Mitigating Degree Bias Adaptively with Hard-to-Learn Nodes in Graph Contrastive Learning

Graph Neural Networks (GNNs) often suffer from degree bias in node classification tasks, where prediction performance varies across nodes with different degrees. Several approaches, which adopt Graph Contrastive Learning (GCL), have been proposed to mitigate this bias. However, the limited number of positive pairs and the equal weighting of all positives and negatives in GCL still lead to low-degree nodes acquiring insufficient and noisy information. This paper proposes the Hardness Adaptive Reweighted (HAR) contrastive loss to mitigate degree bias. It adds more positive pairs by leveraging node labels and adaptively weights positive and negative pairs based on their learning hardness. In addition, we develop an experimental framework named SHARP to extend HAR to a broader range of scenarios. Both our theoretical analysis and experiments validate the effectiveness of SHARP. The experimental results across four datasets show that SHARP achieves better performance against baselines at both global and degree levels.

cs.LG↗

arXiv2024

ProxiMix: Enhancing Fairness with Proximity Samples in Subgroups

Many bias mitigation methods have been developed for addressing fairness issues in machine learning. We found that using linear mixup alone, a data augmentation technique, for bias mitigation, can still retain biases present in dataset labels. Research presented in this paper aims to address this issue by proposing a novel pre-processing strategy in which both an existing mixup method and our new bias mitigation algorithm can be utilized to improve the generation of labels of augmented samples, which are proximity aware. Specifically, we proposed ProxiMix which keeps both pairwise and proximity relationships for fairer data augmentation. We conducted thorough experiments with three datasets, three ML models, and different hyperparameters settings. Our experimental results showed the effectiveness of ProxiMix from both fairness of predictions and fairness of recourse perspectives.

cs.LG↗

arXiv2023

Binomial pmf among arithmetic progressions and sieved sets in random walks

We consider the distribution of the binomial probability mass function (pmf) among arithmetic progressions and obtain an average-type theorem. As applications, we consider the possible visits to a kind of sieved sets of integers or lattice points, by an $\alpha$-random walker. We show that, almost surely, the asymptotic proportion of time the random walker in a sieved set of the type is equal to the density of the set, independently of $\alpha$.

math.NT↗

arXiv2023

Instabilities of a Bose-Einstein condensate with mixed nonlinear and linear lattices

Bose-Einstein condensates (BECs) in periodic potentials generate interesting physics on the instabilities of Bloch states. The lowest-energy Bloch states of BECs in pure nonlinear lattices are dynamically and Landau unstable, which breaks down BEC superfluidity. In this paper we propose to use an out-of-phase linear lattice to stabilize them. The stabilization mechanism is revealed by the averaged interaction. We further incorporate a constant interaction into BECs with mixed nonlinear and linear lattices, and reveal its effect on the instabilities of Bloch states in the lowest band.

cond-mat.quant-gas↗

arXiv2022

Progressively refined deep joint registration segmentation (ProRSeg) of gastrointestinal organs at risk: Application to MRI and cone-beam CT

Method: ProRSeg was trained using 5-fold cross-validation with 110 T2-weighted MRI acquired at 5 treatment fractions from 10 different patients, taking care that same patient scans were not placed in training and testing folds. Segmentation accuracy was measured using Dice similarity coefficient (DSC) and Hausdorff distance at 95th percentile (HD95). Registration consistency was measured using coefficient of variation (CV) in displacement of OARs. Ablation tests and accuracy comparisons against multiple methods were done. Finally, applicability of ProRSeg to segment cone-beam CT (CBCT) scans was evaluated on 80 scans using 5-fold cross-validation. Results: ProRSeg processed 3D volumes (128 $\times$ 192 $\times$ 128) in 3 secs on a NVIDIA Tesla V100 GPU. It's segmentations were significantly more accurate ($p<0.001$) than compared methods, achieving a DSC of 0.94 $\pm$0.02 for liver, 0.88$\pm$0.04 for large bowel, 0.78$\pm$0.03 for small bowel and 0.82$\pm$0.04 for stomach-duodenum from MRI. ProRSeg achieved a DSC of 0.72$\pm$0.01 for small bowel and 0.76$\pm$0.03 for stomach-duodenum from CBCT. ProRSeg registrations resulted in the lowest CV in displacement (stomach-duodenum $CV_{x}$: 0.75\%, $CV_{y}$: 0.73\%, and $CV_{z}$: 0.81\%; small bowel $CV_{x}$: 0.80\%, $CV_{y}$: 0.80\%, and $CV_{z}$: 0.68\%; large bowel $CV_{x}$: 0.71\%, $CV_{y}$ : 0.81\%, and $CV_{z}$: 0.75\%). ProRSeg based dose accumulation accounting for intra-fraction (pre-treatment to post-treatment MRI scan) and inter-fraction motion showed that the organ dose constraints were violated in 4 patients for stomach-duodenum and for 3 patients for small bowel. Study limitations include lack of independent testing and ground truth phantom datasets to measure dose accumulation accuracy.

eess.IV↗

arXiv2022

Ego-graph Replay based Continual Learning for Misinformation Engagement Prediction

Online social network platforms have a problem with misinformation. One popular way of addressing this problem is via the use of machine learning based automated misinformation detection systems to classify if a post is misinformation. Instead of post hoc detection, we propose to predict if a user will engage with misinformation in advance and design an effective graph neural network classifier based on ego-graphs for this task. However, social networks are highly dynamic, reflecting continual changes in user behaviour, as well as the content being posted. This is problematic for machine learning models which are typically trained on a static training dataset, and can thus become outdated when the social network changes. Inspired by the success of continual learning on such problems, we propose an ego-graphs replay strategy in continual learning (EgoCL) using graph neural networks to effectively address this issue. We have evaluated the performance of our method on user engagement with misinformation on two Twitter datasets across nineteen misinformation and conspiracy topics. Our experimental results show that our approach EgoCL has better performance in terms of predictive accuracy and computational resources than the state of the art.

cs.SI↗

arXiv2021

Social Influence Prediction with Train and Test Time Augmentation for Graph Neural Networks

Data augmentation has been widely used in machine learning for natural language processing and computer vision tasks to improve model performance. However, little research has studied data augmentation on graph neural networks, particularly using augmentation at both train- and test-time. Inspired by the success of augmentation in other domains, we have designed a method for social influence prediction using graph neural networks with train- and test-time augmentation, which can effectively generate multiple augmented graphs for social networks by utilising a variational graph autoencoder in both scenarios. We have evaluated the performance of our method on predicting user influence on multiple social network datasets. Our experimental results show that our end-to-end approach, which jointly trains a graph autoencoder and social influence behaviour classification network, can outperform state-of-the-art approaches, demonstrating the effectiveness of train- and test-time augmentation on graph neural networks for social influence prediction. We observe that this is particularly effective on smaller graphs.

cs.SI↗

arXiv2017

Repeat-pass SAR Interferometry Experiments with Gaofen-3: A Case Study of Ningbo Area

This paper reports the repeat-pass interferometric SAR results of Gaofen-3, a Chinese civil SAR satellite, acquired in November 2016 and March 2017 from Ningbo area. With the spatial baseline about 600 m and time baseline 116 days, the coherence of the two images still achieve good enough to generate the digital elevation model (DEM). During the InSAR processing, we compared several baseline estimating methods and obtained a good flat-earth phase removed interferogram map. By using the latest SAR interferogram filter and phase unwrapping method we proposed, we improved the coherence up to 0.88 in urban area and obtained a high quality DEM in Ningbo area. In addition, we evaluated the elevation model by comparing with the elevation values extracted from SRTM. And the result shows that accuracy of the elevation map is about 5 m (RMS) in plane area and 22m (RMS) in mountainous region, which demonstrated that Gaofen-3 has the powerful ability of repeat-pass SAR Interferometry.

physics.geo-ph↗

arXiv2017

Tweet for Behavior Change: Using Social Media for the Dissemination of Public Health Messages

Background: Social media public health campaigns have the advantage of tailored messaging at low cost and large reach, but little is known about what would determine their feasibility as tools for inducing attitude and behavior change. Objective: The aim of this study was to test the feasibility of designing, implementing, and evaluating a social media-enabled intervention for skin cancer prevention. Conclusions: Social media-disseminated public health messages reached more than 23% of the Northern Ireland population. A Web-based survey suggested that the campaign might have contributed to improved knowledge and attitudes toward skin cancer among the target population. Findings suggested that shocking and humorous messages generated greatest impressions and engagement, but information-based messages were likely to be shared most. The extent of behavioral change as a result of the campaign remains to be explored, however, the change of attitudes and knowledge is promising. Social media is an inexpensive, effective method for delivering public health messages. However, existing and traditional process evaluation methods may not be suitable for social media.

cs.SI↗

arXiv2017

Location Inference from Tweets using Grid-based Classification

The impact of social media and its growing association with the sharing of ideas and propagation of messages remains vital in everyday communication. Twitter is one effective platform for the dissemination of news and stories about recent events happening around the world. It has a continually growing database currently adopted by over 300 million users. In this paper we propose a novel grid-based approach employing supervised Multinomial Naive Bayes while extracting geographic entities from relevant user descriptions metadata which gives a spatial indication of the user location. To the best of our knowledge our approach is the first to make location inference from tweets using geo-enriched grid-based classification. Our approach performs better than existing baselines achieving more than 57% accuracy at city-level granularity. In addition we present a novel framework for content-based estimation of user locations by specifying levels of granularity required in pre-defined location grids.

cs.IR↗

arXiv2017

Feasibility Study of Social Media for Public Health Behaviour Changes

Social networking sites such as Twitter and Facebook have been shown to function as effective social sensors that can "feel the pulse" of a community. The aim of the current study is to test the feasibility of designing, implementing and evaluating a bespoke social media-enabled intervention that can be effective for sharing and changing knowledge, attitudes and behaviours in meaningful ways to promote public health, specifically with regards to prevention of skin cancer. We present the design and implementation details of the campaign followed by summary findings and analysis.

cs.CY↗

arXiv2017

A survey of location inference techniques on Twitter

The increasing popularity of the social networking service, Twitter, has made it more involved in day-to-day communications, strengthening social relationships and information dissemination. Conversations on Twitter are now being explored as indicators within early warning systems to alert of imminent natural disasters such earthquakes and aid prompt emergency responses to crime. Producers are privileged to have limitless access to market perception from consumer comments on social media and microblogs. Targeted advertising can be made more effective based on user profile information such as demography, interests and location. While these applications have proven beneficial, the ability to effectively infer the location of Twitter users has even more immense value. However, accurately identifying where a message originated from or author's location remains a challenge thus essentially driving research in that regard. In this paper, we survey a range of techniques applied to infer the location of Twitter users from inception to state-of-the-art. We find significant improvements over time in the granularity levels and better accuracy with results driven by refinements to algorithms and inclusion of more spatial features.

cs.SI↗

arXiv2013

Subthreshold pion production within a transport description of central Au+Au collisions

Mapping out the equation of state (EOS) of nuclear matter is a long standing problem in nuclear physics. Recent emphasis is onto the density dependence of the symmetry energy, with experiments needing dedicated symmetry-energy observables. Towards the latter goal, we employ pBUU transport model to simulate pion production in heavy ion collision (HIC). We find that the net pion yield tests the momentum dependence of nuclear mean field (MF). In exploring the sensitivity of pion observables to the symmetry energy at higher than normal densities, we find that our calculations of pion ratios contradict, at some level, predictions from both IBUU and ImIQMD models. We propose to employ the pion ratio in the high-energy tail of spectra in future experiments, to distinguish between different variants of high-density symmetry energy.

nucl-th↗