Fujun Liu

Peiqi Wang, ShengYun Peng, Xuewen Zhang et al.

This work investigates the optimal allocation of inference compute across three key scaling factors in video vision language models: language model size, frame count, and the number of visual tokens per frame. While prior works typically focuses on optimizing model efficiency or improving performance without considering resource constraints, we instead identify optimal model configuration under fixed inference compute budgets. We conduct large-scale training sweeps and careful parametric modeling of task performance to identify the inference compute-optimal frontier. Our experiments reveal how task performance depends on scaling factors and finetuning data size, as well as how changes in data size shift the compute-optimal frontier. These findings translate to practical tips for selecting these scaling factors.

Ze Tao, Darui Zhao, Fujun Liu et al.

Physics-informed learning for PDEs is surging across scientific computing and industrial simulation, yet prevailing methods face spectral bias, residual-data imbalance, and weak extrapolation. We introduce a representation-level spectral remodeling xLSTM-PINN that combines gated-memory multiscale feature extraction with adaptive residual-data weighting to curb spectral bias and strengthen extrapolation. Across four benchmarks, we integrate gated cross-scale memory, a staged frequency curriculum, and adaptive residual reweighting, and verify with analytic references and extrapolation tests, achieving markedly lower spectral error and RMSE and a broader stable learning-rate window. Frequency-domain benchmarks show raised high-frequency kernel weights and a right-shifted resolvable bandwidth, shorter high-k error decay and time-to-threshold, and narrower error bands with lower MSE, RMSE, MAE, and MaxAE. Compared with the baseline PINN, we reduce MSE, RMSE, MAE, and MaxAE across all four benchmarks and deliver cleaner boundary transitions with attenuated high-frequency ripples in both frequency and field maps. This work suppresses spectral bias, widens the resolvable band and shortens the high-k time-to-threshold under the same budget, and without altering AD or physics losses improves accuracy, reproducibility, and transferability.

Fu-jun Liu, Zhong-qing Wang, Huiyuan Li

A fully diagonalized spectral method using generalized Laguerre functions is proposed and analyzed for solving elliptic equations on the half line. We first define the generalized Laguerre functions which are complete and mutually orthogonal with respect to an equivalent Sobolev inner product. Then the Fourier-like Sobolev orthogonal basis functions are constructed for the diagonalized Laguerre spectral method of elliptic equations. Besides, a unified orthogonal Laguerre projection is established for various elliptic equations. On the basis of this orthogonal Laguerre projection, we obtain optimal error estimates of the fully diagonalized Laguerre spectral method for both Dirichlet and Robin boundary value problems. Finally, numerical experiments, which are in agreement with the theoretical analysis, demonstrate the effectiveness and the spectral accuracy of our diagonalized method.