arXiv subjects

Aaron Zhang

Publications and source records attributed to Aaron Zhang.

5 recordsLinked to original sources

arXiv2025

EfficientSAM3: Progressive Hierarchical Distillation for Video Concept Segmentation from SAM1, 2, and 3

The Segment Anything Model 3 (SAM3) advances visual understanding with Promptable Concept Segmentation (PCS) across images and videos, but its unified architecture (shared vision backbone, DETR-style detector, dense-memory tracker) remains prohibitive for on-device use. We present EfficientSAM3, a family of efficient models built on Progressive Hierarchical Distillation (PHD) that transfers capability from SAM3 to lightweight students in three stages: (1) Encoder Distillation aligns image features via prompt-in-the-loop training on SA-1B; (2) Temporal Memory Distillation replaces dense memory with a compact Perceiver-based module trained on SA-V to compress and retrieve spatiotemporal features efficiently; and (3) End-to-End Fine-Tuning refines the full pipeline on the official SAM3 PCS data to preserve concept-level performance. PHD yields a spectrum of student variants using RepViT, TinyViT, and EfficientViT backbones, enabling on-device concept segmentation and tracking while maintaining high fidelity to teacher behavior. We benchmark on popular VOS datasets, and compare with varies of releated work, achieing strong performance-efficiency trade-offs.

cs.CV↗

arXiv2025

CULTURE3D: A Large-Scale and Diverse Dataset of Cultural Landmarks and Terrains for Gaussian-Based Scene Rendering

Current state-of-the-art 3D reconstruction models face limitations in building extra-large scale outdoor scenes, primarily due to the lack of sufficiently large-scale and detailed datasets. In this paper, we present a extra-large fine-grained dataset with 10 billion points composed of 41,006 drone-captured high-resolution aerial images, covering 20 diverse and culturally significant scenes from worldwide locations such as Cambridge Uni main buildings, the Pyramids, and the Forbidden City Palace. Compared to existing datasets, ours offers significantly larger scale and higher detail, uniquely suited for fine-grained 3D applications. Each scene contains an accurate spatial layout and comprehensive structural information, supporting detailed 3D reconstruction tasks. By reconstructing environments using these detailed images, our dataset supports multiple applications, including outputs in the widely adopted COLMAP format, establishing a novel benchmark for evaluating state-of-the-art large-scale Gaussian Splatting methods.The dataset's flexibility encourages innovations and supports model plug-ins, paving the way for future 3D breakthroughs. All datasets and code will be open-sourced for community use.

cs.CV↗

arXiv2022

Bounds on the Total Coefficient Size of Nullstellensatz Proofs of the Pigeonhole Principle and the Ordering Principle

In this paper, we investigate the total coefficient size of Nullstellensatz proofs. We show that Nullstellensatz proofs of the pigeonhole principle on $n$ pigeons require total coefficient size $2^{Ω(n)}$ and that there exist Nullstellensatz proofs of the ordering principle on $n$ elements with total coefficient size $2^n - n$.

cs.CC↗

arXiv2019

A Statistical Analysis of Recent Traffic Crashes in Massachusetts

A statistical analysis implemented in the Python programming language was performed on the available MassDOT car accident data to identify whether a certain set of traffic circumstances would increase the likelihood of injuries. In the analysis, we created a binary classifier as a model to separate crashes that resulted in injury from those that did not. To accomplish this, we first cleaned up the initial data, then proceeded to represent categorical variables numerically through one hot encoding before finally producing models with Recursive Feature Elimination (RFE) and without RFE, in conjunction with logistic regression. This statistical analysis plays a significant role in our modern road network that has presented us with a heap of obstacles, one of the most critical being the issue of how we can ensure the safety of all drivers and passengers. Findings from our analysis identify that tough weather and road conditions, senior/teen drivers and dangerous intersections play prominent roles in accidents that resulted in injuries in Massachusetts. These new findings can provide valuable references and scientific data support to relevant authorities and policy makers for upgrading road infrastructure, passing regulations, etc.

stat.AP↗

arXiv2019

Stackelberg Punishment and Bully-Proofing Autonomous Vehicles

Mutually beneficial behavior in repeated games can be enforced via the threat of punishment, as enshrined in game theory's well-known "folk theorem." There is a cost, however, to a player for generating these disincentives. In this work, we seek to minimize this cost by computing a "Stackelberg punishment," in which the player selects a behavior that sufficiently punishes the other player while maximizing its own score under the assumption that the other player will adopt a best response. This idea generalizes the concept of a Stackelberg equilibrium. Known efficient algorithms for computing a Stackelberg equilibrium can be adapted to efficiently produce a Stackelberg punishment. We demonstrate an application of this idea in an experiment involving a virtual autonomous vehicle and human participants. We find that a self-driving car with a Stackelberg punishment policy discourages human drivers from bullying in a driving scenario requiring social negotiation.

cs.HC↗