Selected work

Representative Papers

EfficientPENet: Real-Time Depth Completion from Sparse LiDAR via Lightweight Multi-Modal Fusion

Apr 20, 2026

This work addresses the challenge of efficient depth completion from sparse LiDAR point clouds and RGB images by proposing a lightweight dual-branch network. Built upon ConvNeXt as the backbone, the method integrates sparsity-invariant convolutions with a Convolutional Spatial Propagation Network (CSPN), employing late fusion and multi-scale depth supervision. It is the first to introduce layer normalization, large-kernel depthwise convolutions, and stochastic depth regularization into depth completion. Additionally, a position-aware test-time augmentation strategy is devised. On the KITTI benchmark, the model achieves an RMSE of 631.94 mm with only 36.24M parameters and a latency of 20.51 ms (48.76 FPS), reducing the parameter count by 3.7× and accelerating inference by 23× compared to BP-Net.

0 citationsRead paper

RADAR: Exposing Unlogged NoSQL Operations

Feb 13, 2026

The widespread adoption of NoSQL databases has made digital forensics increasingly difficult as storage formats are diverse and often opaque, and audit logs cannot be assumed trustworthy when privileged insiders, such as DevOps or administrators, can disable, suppress, or manipulate logging to conceal activity. We present RADAR (Record&Artifact Detection, Alignment&Reporting), a log-adversary-aware framework that derives forensic ground truth by cross-referencing low-level storage artifacts against high-level application logs. RADAR analyzes artifacts reconstructed by the Automated NoSQL Carver (ANOC), which infers layouts and carves records directly from raw disk bytes, bypassing database APIs and the management system entirely, thereby treating physical storage as the independent evidence source. RADAR then reconciles carved artifacts with the audit log to identify delta artifacts such as unlogged insertions, silent deletions, and field-level updates that exist on disk but are absent from the logical history. We evaluate RADAR across ten NoSQL engines, including BerkeleyDB, LMDB, MDBX, etcd, ZODB, Durus, LiteDB, Realm, RavenDB, and NitriteDB, spanning key-value and document stores and multiple storage designs, e.g., copy-on-write/MVCC, B/B+ tree, and append-only. Under log-evasion scenarios, such as log suppression and post-maintenance attacks, including cases where historical bytes are pruned, RADAR consistently exposes unattributed operations while sustaining 31.7-397 MB/min processing throughput, demonstrating the feasibility of log-independent, trustworthy NoSQL forensics.

0 citationsRead paper

Edge-Optimized Vision-Language Models for Underground Infrastructure Assessment

Feb 03, 2026

This work addresses the challenge of automatically generating human-readable defect summaries from visual inspections of underground infrastructure—such as drainage pipes and culverts—on resource-constrained edge devices. The authors propose a two-stage, end-to-end lightweight pipeline: first, an efficient defect segmentation model, RAPID-SCAN, with only 0.64M parameters achieves an F1-score of 0.834; second, a quantized and fine-tuned Phi-3.5 vision-language model produces domain-specific natural language summaries. This study presents the first integration of lightweight segmentation with an edge-optimized vision-language model, introduces the first dedicated dataset with human-verified descriptive annotations, and demonstrates real-time inference on a mobile robotic platform, significantly enhancing both the interpretability and deployment efficiency of intelligent infrastructure assessment systems.

0 citationsRead paper

Reinforcement Learning in Financial Decision Making: A Systematic Review of Performance, Challenges, and Implementation Strategies

Dec 11, 2025

Reinforcement learning (RL) faces critical challenges in financial decision-making, including poor interpretability, limited robustness, deployment difficulties, regulatory compliance hurdles, weak adaptability to non-stationary market environments, and the absence of standardized benchmarks. This study conducts a systematic review of RL applications in market making, portfolio optimization, and algorithmic trading from 2017–2025, performing a meta-analysis of 167 peer-reviewed works. Integrating insights from financial microstructure theory, regulatory constraints, and risk management—augmented by synthetic-data-driven empirical validation—the work proposes a novel unified framework that explicitly embeds domain knowledge into architectural design, prioritizing implementation quality and model interpretability over algorithmic complexity. Results demonstrate that RL significantly outperforms conventional approaches in market-making tasks; however, regulatory-compliant, interpretable models are urgently needed, alongside cross-task, reproducible, standardized evaluation protocols to bridge the gap between research and real-world deployment.

0 citationsRead paper

Detection of Anomalous Behavior in Robot Systems Based on Machine Learning

Sep 12, 2025

This study addresses anomaly detection in robotic systems to enhance operational safety and reliability. We propose a context-aware machine learning approach, leveraging system logs collected from quadcopter and Pioneer robots under diverse operational conditions in the CoppeliaSim simulation environment. We systematically evaluate the anomaly detection performance of logistic regression (LR), support vector machines (SVM), and autoencoders. Our key contribution is a “context-dependent model selection strategy,” which empirically demonstrates that anomaly complexity varies significantly across robotic platforms: LR achieves optimal performance in highly dynamic quadcopter scenarios, whereas autoencoders substantially outperform other models on the less structured, behaviorally complex Pioneer platform. Experimental results validate the adaptability and effectiveness of this strategy across heterogeneous robotic systems, providing a principled methodology for tailoring anomaly detection solutions to specific robot architectures and operational contexts.

0 citationsRead paper

Recent publications

Latest Papers

EfficientPENet: Real-Time Depth Completion from Sparse LiDAR via Lightweight Multi-Modal Fusion

Apr 20, 2026

This work addresses the challenge of efficient depth completion from sparse LiDAR point clouds and RGB images by proposing a lightweight dual-branch network. Built upon ConvNeXt as the backbone, the method integrates sparsity-invariant convolutions with a Convolutional Spatial Propagation Network (CSPN), employing late fusion and multi-scale depth supervision. It is the first to introduce layer normalization, large-kernel depthwise convolutions, and stochastic depth regularization into depth completion. Additionally, a position-aware test-time augmentation strategy is devised. On the KITTI benchmark, the model achieves an RMSE of 631.94 mm with only 36.24M parameters and a latency of 20.51 ms (48.76 FPS), reducing the parameter count by 3.7× and accelerating inference by 23× compared to BP-Net.

0 citationsRead paper

RADAR: Exposing Unlogged NoSQL Operations

Feb 13, 2026

The widespread adoption of NoSQL databases has made digital forensics increasingly difficult as storage formats are diverse and often opaque, and audit logs cannot be assumed trustworthy when privileged insiders, such as DevOps or administrators, can disable, suppress, or manipulate logging to conceal activity. We present RADAR (Record&Artifact Detection, Alignment&Reporting), a log-adversary-aware framework that derives forensic ground truth by cross-referencing low-level storage artifacts against high-level application logs. RADAR analyzes artifacts reconstructed by the Automated NoSQL Carver (ANOC), which infers layouts and carves records directly from raw disk bytes, bypassing database APIs and the management system entirely, thereby treating physical storage as the independent evidence source. RADAR then reconciles carved artifacts with the audit log to identify delta artifacts such as unlogged insertions, silent deletions, and field-level updates that exist on disk but are absent from the logical history. We evaluate RADAR across ten NoSQL engines, including BerkeleyDB, LMDB, MDBX, etcd, ZODB, Durus, LiteDB, Realm, RavenDB, and NitriteDB, spanning key-value and document stores and multiple storage designs, e.g., copy-on-write/MVCC, B/B+ tree, and append-only. Under log-evasion scenarios, such as log suppression and post-maintenance attacks, including cases where historical bytes are pruned, RADAR consistently exposes unattributed operations while sustaining 31.7-397 MB/min processing throughput, demonstrating the feasibility of log-independent, trustworthy NoSQL forensics.

0 citationsRead paper

Edge-Optimized Vision-Language Models for Underground Infrastructure Assessment

Feb 03, 2026

This work addresses the challenge of automatically generating human-readable defect summaries from visual inspections of underground infrastructure—such as drainage pipes and culverts—on resource-constrained edge devices. The authors propose a two-stage, end-to-end lightweight pipeline: first, an efficient defect segmentation model, RAPID-SCAN, with only 0.64M parameters achieves an F1-score of 0.834; second, a quantized and fine-tuned Phi-3.5 vision-language model produces domain-specific natural language summaries. This study presents the first integration of lightweight segmentation with an edge-optimized vision-language model, introduces the first dedicated dataset with human-verified descriptive annotations, and demonstrates real-time inference on a mobile robotic platform, significantly enhancing both the interpretability and deployment efficiency of intelligent infrastructure assessment systems.

0 citationsRead paper

Reinforcement Learning in Financial Decision Making: A Systematic Review of Performance, Challenges, and Implementation Strategies

Dec 11, 2025

Reinforcement learning (RL) faces critical challenges in financial decision-making, including poor interpretability, limited robustness, deployment difficulties, regulatory compliance hurdles, weak adaptability to non-stationary market environments, and the absence of standardized benchmarks. This study conducts a systematic review of RL applications in market making, portfolio optimization, and algorithmic trading from 2017–2025, performing a meta-analysis of 167 peer-reviewed works. Integrating insights from financial microstructure theory, regulatory constraints, and risk management—augmented by synthetic-data-driven empirical validation—the work proposes a novel unified framework that explicitly embeds domain knowledge into architectural design, prioritizing implementation quality and model interpretability over algorithmic complexity. Results demonstrate that RL significantly outperforms conventional approaches in market-making tasks; however, regulatory-compliant, interpretable models are urgently needed, alongside cross-task, reproducible, standardized evaluation protocols to bridge the gap between research and real-world deployment.

0 citationsRead paper

Detection of Anomalous Behavior in Robot Systems Based on Machine Learning

Sep 12, 2025

This study addresses anomaly detection in robotic systems to enhance operational safety and reliability. We propose a context-aware machine learning approach, leveraging system logs collected from quadcopter and Pioneer robots under diverse operational conditions in the CoppeliaSim simulation environment. We systematically evaluate the anomaly detection performance of logistic regression (LR), support vector machines (SVM), and autoencoders. Our key contribution is a “context-dependent model selection strategy,” which empirically demonstrates that anomaly complexity varies significantly across robotic platforms: LR achieves optimal performance in highly dynamic quadcopter scenarios, whereas autoencoders substantially outperform other models on the less structured, behaviorally complex Pioneer platform. Experimental results validate the adaptability and effectiveness of this strategy across heterogeneous robotic systems, providing a principled methodology for tailoring anomaly detection solutions to specific robot architectures and operational contexts.

0 citationsRead paper