More Like this

1. Deep neural network for solving Poisson-Boltzmann equations on protein surfaces

   https://doi.org/10.1615/JMachLearnModelComput.2025057332  

   Chen, Duan; He, Cuiyu; Hu, Xiaozhe; Mu, Lin; Shen, Zhaiming (January 2025, Journal of Machine Learning for Modeling and Computing)

   In this paper, we propose a new deep learning method for the nonlinear Poisson-Boltzmann problems with applications in computational biology. To tackle the discontinuity of the solution, e.g., across protein surfaces, we approximate the solution by a piecewise mesh-free neural network that can capture the dramatic change in the solution across the interface. The partial differential equation problem is first reformulated as a least-squares physics-informed neural network (PINN)-type problem and then discretized to an objective function using mean squared error via sampling. The solution is obtained by minimizing the designed objective function via standard training algorithms such as the stochastic gradient descent method. Finally, the effectiveness and efficiency of the neural network are validated using complex protein interfaces on various manufactured functions with different frequencies. 

   more » « less
2. Improved Parallel Construction of Wavelet Trees and Rank/Select Structures

   Shun, Julian (January 2020, Information and computation)

   Existing parallel algorithms for wavelet tree construction have a work complexity of $$O(n\log\sigma)$$. This paper presents parallel algorithms for the problem with improved work complexity. Our first algorithm is based on parallel integer sorting and has either $$O(n\log\log n\lceil\log\sigma/\sqrt{\log n\log\log n}\rceil)$$ work and polylogarithmic depth, or $$O(n\lceil\log\sigma/\sqrt{\log n}\rceil)$$ work and sub-linear depth. We also describe another algorithm that has $$O(n\lceil\log\sigma/\sqrt{\log n} \rceil)$$ work and $$O(\sigma+\log n)$$ depth. We then show how to use similar ideas to construct variants of wavelet trees (arbitrary-shaped binary trees and multiary trees) as well as wavelet matrices in parallel with lower work complexity than prior algorithms. Finally, we show that the rank and select structures on binary sequences and multiary sequences, which are stored on wavelet tree nodes, can be constructed in parallel with improved work bounds, matching those of the best existing sequential algorithms for constructing rank and select structures. 

   more » « less
3. Affine Stresses, Inverse Systems, and Reconstruction Problems

   https://doi.org/10.1093/imrn/rnad294  

   Murai, Satoshi; Novik, Isabella; Zheng, Hailun (December 2023, International Mathematics Research Notices)

   Abstract A conjecture of Kalai asserts that for $$d\geq 4$$, the affine type of a prime simplicial $$d$$-polytope $$P$$ can be reconstructed from the space of affine $$2$$-stresses of $$P$$. We prove this conjecture for all $$d\geq 5$$. We also prove the following generalization: for all pairs $(i,d)$ with $$2\leq i\leq \lceil \frac d 2\rceil -1$$, the affine type of a simplicial $$d$$-polytope $$P$$ that has no missing faces of dimension $$\geq d-i+1$$ can be reconstructed from the space of affine $$i$$-stresses of $$P$$. A consequence of our proofs is a strengthening of the Generalized Lower Bound Theorem: it was proved by Nagel that for any simplicial $(d-1)$-sphere $$\Delta $$ and $$1\leq k\leq \lceil \frac {d}{2}\rceil -1$$, $$g_{k}(\Delta )$$ is at least as large as the number of missing $(d-k)$-faces of $$\Delta $$; here we show that, for $$1\leq k\leq \lfloor \frac {d}{2}\rfloor -1$$, equality holds if and only if $$\Delta $$ is $$k$$-stacked. Finally, we show that for $$d\geq 4$$, any simplicial $$d$$-polytope $$P$$ that has no missing faces of dimension $$\geq d-1$$ is redundantly rigid, that is, for each edge $$e$$ of $$P$$, there exists an affine $$2$$-stress on $$P$$ with a non-zero value on $$e$$. 

   more » « less
4. Deep Network Approximation: Achieving Arbitrary Accuracy with Fixed Number of Neurons

   Shijun Zhang; Zuowei Shen; Haizhao Yang (January 2022, Journal of machine learning research)

   Ruslan Salakhutdinov (Ed.)

   This paper develops simple feed-forward neural networks that achieve the universal approximation property for all continuous functions with a fixed finite number of neurons. These neural networks are simple because they are designed with a simple and computable continuous activation function $$\sigma$$ leveraging a triangular-wave function and a softsign function. We prove that $$\sigma$$-activated networks with width $36d(2d+1)$ and depth $11$ can approximate any continuous function on a $$d$$-dimensional hypercube within an arbitrarily small error. Hence, for supervised learning and its related regression problems, the hypothesis space generated by these networks with a size not smaller than $$36d(2d+1)\times 11$$ is dense in the space of continuous functions. Furthermore, classification functions arising from image and signal classification are in the hypothesis space generated by $$\sigma$$-activated networks with width $36d(2d+1)$ and depth $12$, when there exist pairwise disjoint closed bounded subsets of $$\mathbb{R}^d$$ such that the samples of the same class are located in the same subset. 

   more » « less
5. Deep Learning meets Nonparametric Regression: Are Weight-Decayed DNNs Locally Adaptive?

   Zhang, Kaiqi; Wang, Yu-Xiang (February 2023, International Conference on Learning Representation (ICLR-23))

   We study the theory of neural network (NN) from the lens of classical nonparametric regression problems with a focus on NN's ability to adaptively estimate functions with heterogeneous smoothness -- a property of functions in Besov or Bounded Variation (BV) classes. Existing work on this problem requires tuning the NN architecture based on the function spaces and sample sizes. We consider a "Parallel NN" variant of deep ReLU networks and show that the standard weight decay is equivalent to promoting the ℓp-sparsity (0<1) of the coefficient vector of an end-to-end learned function bases, i.e., a dictionary. Using this equivalence, we further establish that by tuning only the weight decay, such Parallel NN achieves an estimation error arbitrarily close to the minimax rates for both the Besov and BV classes. Notably, it gets exponentially closer to minimax optimal as the NN gets deeper. Our research sheds new lights on why depth matters and how NNs are more powerful than kernel methods. 

   more » « less