arXiv subjects

Xiaohong Deng

Publications and source records attributed to Xiaohong Deng.

3 recordsLinked to original sources

arXiv2026

MLFCIL: A Multi-Level Forgetting Mitigation Framework for Federated Class-Incremental Learning in LEO Satellites

Low-Earth-orbit (LEO) satellite constellations are increasingly performing on-board computing. However, the continuous emergence of new classes under strict memory and communication constraints poses major challenges for collaborative training. Federated class-incremental learning (FCIL) enables distributed incremental learning without sharing raw data, but faces three LEO-specific challenges: non-independent and identically distributed data heterogeneity caused by orbital dynamics, amplified catastrophic forgetting during aggregation, and the need to balance stability and plasticity under limited resources. To tackle these challenges, we propose MLFCIL, a multi-level forgetting mitigation framework that decomposes catastrophic forgetting into three sources and addresses them at different levels: class-reweighted loss to reduce local bias, knowledge distillation with feature replay and prototype-guided drift compensation to preserve cross-task knowledge, and class-aware aggregation to mitigate forgetting during federation. In addition, we design a dual-granularity coordination strategy that combines round-level adaptive loss balancing with step-level gradient projection to further enhance the stability-plasticity trade-off. Experiments on the NWPU-RESISC45 dataset show that MLFCIL significantly outperforms baselines in both accuracy and forgetting mitigation, while introducing minimal resource overhead.

cs.NI↗

arXiv2021

Comparing Broadband ISP Performance using Big Data from M-Lab

Comparing ISPs on broadband speed is challenging, since measurements can vary due to subscriber attributes such as operation system and test conditions such as access capacity, server distance, TCP window size, time-of-day, and network segment size. In this paper, we draw inspiration from observational studies in medicine, which face a similar challenge in comparing the effect of treatments on patients with diverse characteristics, and have successfully tackled this using "causal inference" techniques for {\em post facto} analysis of medical records. Our first contribution is to develop a tool to pre-process and visualize the millions of data points in M-Lab at various time- and space-granularities to get preliminary insights on factors affecting broadband performance. Next, we analyze 24 months of data pertaining to twelve ISPs across three countries, and demonstrate that there is observational bias in the data due to disparities amongst ISPs in their attribute distributions. For our third contribution, we apply a multi-variate matching method to identify suitable cohorts that can be compared without bias, which reveals that ISPs are closer in performance than thought before. Our final contribution is to refine our model by developing a method for estimating speed-tier and re-apply matching for comparison of ISP performance. Our results challenge conventional rankings of ISPs, and pave the way towards data-driven approaches for unbiased comparisons of ISPs world-wide.

cs.PF↗

arXiv2019

Estimating Residential Broadband Capacity using Big Data from M-Lab

Knowing residential broadband capacity profiles across a population is of interest to both consumers and regulators who want to compare or audit performance of various broadband service offerings. Unfortunately, extracting broadband capacity from speed tests in public datasets like M-Lab is challenging because tests are indexed by client IP address which can be dynamic and/or obfuscated by NAT, and variable network conditions can affect measurements. This paper presents the first systematic effort to isolate households and extract their broadband capacity using 63 million speed test measurements recorded over a 12 month period in the M-Lab dataset. We first identify a key parameter, the correlation between measured speed and congestion count for a specific client IP address, as an indicator of whether the IP address represents a single house, or a plurality of houses that may be dynamically sharing addresses or be aggregated behind a NAT. We then validate our approach by comparing to ground truth taken from a few known houses, and at larger scale by checking internal consistency across ISPs and across months. Lastly, we present results that isolate households and estimate their broadband capacity based on measured data, and additionally reveal insights into the prevalence of NAT and variations in service capacity tiers across ISPs.

cs.NI↗