Yidi Zhang

Hao Zhu, Haotian Yang, Longwei Guo et al.

In this paper, we present a large-scale detailed 3D face dataset, FaceScape, and the corresponding benchmark to evaluate single-view facial 3D reconstruction. By training on FaceScape data, a novel algorithm is proposed to predict elaborate riggable 3D face models from a single image input. FaceScape dataset releases $16,940$ textured 3D faces, captured from $847$ subjects and each with $20$ specific expressions. The 3D models contain the pore-level facial geometry that is also processed to be topologically uniform. These fine 3D facial models can be represented as a 3D morphable model for coarse shapes and displacement maps for detailed geometry. Taking advantage of the large-scale and high-accuracy dataset, a novel algorithm is further proposed to learn the expression-specific dynamic details using a deep neural network. The learned relationship serves as the foundation of our 3D face prediction system from a single image input. Different from most previous methods, our predicted 3D models are riggable with highly detailed geometry under different expressions. We also use FaceScape data to generate the in-the-wild and in-the-lab benchmark to evaluate recent methods of single-view face reconstruction. The accuracy is reported and analyzed on the dimensions of camera pose and focal length, which provides a faithful and comprehensive evaluation and reveals new challenges. The unprecedented dataset, benchmark, and code have been released at https://github.com/zhuhao-nju/facescape.

Yash Jangir, Yidi Zhang, Kashu Yamazaki et al.

The pursuit of robot generalists - instructable agents capable of performing diverse tasks across diverse environments - demands rigorous and scalable evaluation. Yet real-world testing of robot policies remains fundamentally constrained: it is labor-intensive, slow, unsafe at scale, and difficult to reproduce. Existing simulation benchmarks are similarly limited, as they train and test policies within the same synthetic domains and cannot assess models trained from real-world demonstrations or alternative simulation environments. As policies expand in scope and complexity, these barriers only intensify, since defining "success" in robotics often hinges on nuanced human judgments of execution quality. In this paper, we introduce a new benchmarking framework that overcomes these challenges by shifting VLA evaluation into large-scale simulated environments augmented with online human feedback. Leveraging advances in vision-language models, 2D-to-3D generative modeling, and differentiable rendering, our approach automatically converts video demonstrations from widely used robot datasets into simulated counterparts. Within these digital twins, we assess VLA policies using both automated VLM-guided scoring and scalable human preference judgments collected from crowdworkers, transforming human involvement from tedious scene setup, resetting, and safety supervision into lightweight preference comparisons. To measure robustness, we systematically perturb simulated environments along multiple axes, such as textures and object placements, stress-testing policy generalization under controlled variation. The result is a continuously evolving, reproducible, and scalable benchmark for real-world trained robot manipulation policies, addressing a critical missing capability in today's robotics landscape.

Jindi Wang, Yidi Zhang, Zhaoxing Li

This study presents DeBERTa-KC, a transformer-based model for automatic classification of knowledge construction (KC) levels in online science learning discourse. Using comments collected from four popular YouTube science channels (2022--2024), a balanced corpus of 20,000 manually annotated samples was created across four KC categories: \textit{nonKC}, \textit{Share}, \textit{Explore}, and \textit{Negotiate}. The proposed model extends DeBERTa-v3 with Focal Loss, Label Smoothing, and R-Drop regularization to address class imbalance and enhance generalization. A reproducible end-to-end pipeline was implemented, encompassing data extraction, annotation, preprocessing, training, and evaluation. Across 10-fold stratified cross-validation, DeBERTa-KC achieved a macro-F1 of $0.836 \pm 0.008$, significantly out-performing both classical and transformer baselines ($p<0.01$). Per-category results indicate strong sensitivity to higher-order epistemic engagement, particularly in \textit{Explore} and \textit{Negotiate} discourse. These findings demonstrate that large language models can effectively capture nuanced indicators of knowledge construction in informal digital learning environments, offering scalable, theory-informed approaches to discourse analysis and the development of automated tools for assessing epistemic engagement.