Street-view imagery is increasingly analysed with vision-language models (VLMs) to infer urban attributes, but predictive accuracy alone does not show how much a photograph contributes beyond data already available for the same place. Using three VLMs, we compare image-based predictions with existing urban data for sev...
Street-view imagery is increasingly used to infer urban attributes, but predictive accuracy alone does not reveal how much a photograph contributes beyond data already available for the same place. We compare image-based predictions with existing urban data across seven attributes from five public resources and three V...
The results distinguish cross-sensor consistency, multimodal fusion, and generalization to new action orders as separate questions as separate questions in evaluating multimodal physical representations.
Kai-Zhen Tan, Xin Xu, Si-Ru Tao et al.· 0 citations
Pretraining how a street is appraised is weak evidence that a model represents the street as the brain does, and prediction accuracy and neural correspondence were not reliably related across models, and the best predictor was among the least aligned.
This work builds the flow-sensitive type system family that the quantale calls for, and shows that the object which makes such families useful, the universal type object from which every member specialises, splits in two, with a consequence for enforcement.
EquiSD exploits an exact physical symmetry to provide label-free supervision for improving metric grounding in vision-language models and generalizes to unseen world scales and transfers without adaptation to real QuantiPhy videos, where accuracy increases by 6.4 points.
Kai-Zhen Tan, Yang Feng, He-Qing Du et al.· 0 citations
CacheWeaver is presented, a lightweight prompt-layer method for cache-aware evidence ordering that keeps a prefix tree over recently served evidence sequences and uses a greedy walk to place the most reusable prefix first, while leaving the serving engine and retrieved evidence set unchanged.
Kai-Zhen Tan, Rong Gu, Mingyuan Li· arXiv.org· 1 citation
Vision-language models (VLMs) are increasingly used to measure urban change from repeated street-level imagery, and the results are often mapped for individual sample points. Such maps assume that a perception score changes only if the street does. We test this assumption with repeated Google Street View captures of un...
Extending compression-based memorization analysis to the frozen-base setting, this work measures directly, in bits, how much a low-rank adapter writes into a model it never changes, finding that the answer is both smaller than full fine-tuning and less lawful than parameter counting would predict.
Kaizhen Tan, Heqing Du, Yang Feng· arXiv.org· 0 citations
A central premise of latent world models is that predicting the future forces a representation to internalize the physics of its environment. Which physical quantities does a trained latent actually contain, and what decides this? We answer with controlled interventions in POKEWORLD, an interactive environment whose vi...
Kai-Zhen Tan, Xin Xu, Siru Tao et al.· arXiv.org· 2 citations
Empirical work on algorithmic collusion asks one question of the data: are prices supracompetitive? We show this can be answered"no"by a conspiracy that is nonetheless profitable. Consider bidding agents that couple only through the joint distribution of their unexplained bid components, leaving every agent's own bid l...
Xin Xu, Cheng-Rui Wu, Jiayu Lu et al.· arXiv.org· 1 citation
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
