It is proved that any corrector measurable with respect to past observations leaves at least the conditional variance of the statistic it tracks, and that trend extrapolation beats trusting the last observation only when $\mu^2>v(1-2\rho)$.
Tian-Peng Li, Xuan Guo, Wenjun Wang et al.· arXiv.org· 0 citations
This survey presents the first systematic review of DHG representation learning methods and proposes a novel algorithm-centric taxonomy that categorizes existing literature, including early embedding-based approaches, graph neural network (GNN)-based models, and relatively recent Transformer-based DHG methods, while explicitly highlighting their intrinsic modeling biases with respect to dynamic granularity.
Huan Liu, Peng-Fei Jiao, Jie Yin et al.· Proceedings of the Thirty-Fi...· 0 citations
This work identifies the Logit Conflation Problem, where a to-ken’s logit aggregates prompt-independent factors, including linguistic fluency and parametric associations, with prompt-relevance, and proposes SEAL-Sampling to isolate this component through attention-weighted attribution.
Pinlong Zhao, Huijun Tang, Pengfei Jiao et al.· Annual Meeting of the Associ...· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
