Categorizing white blood cells (WBCs) in blood specimens is crucial for early disease detection in hematology and pathology. Numerous research endeavors have been undertaken to investigate the precise classification of WBCs in microscopic blood smear images, utilizing machine learning and deep learning methodologies. However, the notable accuracy achieved by most of these techniques relies on the assumption that the conditions under which the test microscope images are acquired closely resemble those of the training images, including the use of identical microscopes and consistent lighting conditions. The primary aim of this research is to design a novel network layer that can be used as a component in combination with other backone network, and improve the generalization ability of this backbone for classification tasks on WBC data. In this investigation, we predominantly employ the color invariance technique to fashion a trainable convolutional layer, which serves as the tool for extracting edge features from the original image. This layer is amenable to integration with various backbone architectures. Ultimately, the model was trained utilizing the Raabin-WBC dataset, and its efficacy was evaluated by appraising its performance on WBC images acquired through a variety of microscope instruments.

Representation learning based on dynamic graphs has received a lot of attention in recent years due to its wide range of application scenarios. Although many discrete or continuous dynamic graph representation learning methods have been proposed, many of them ignore the role of edge types. Through the observation of dynamic graphs in the real world, it is found that the types of various edges are very different in nature. They are roughly divided into two categories according to the frequency of occurrence: evolutive edges that appear infrequently and interactive edges that appear frequently. For both types of edges, we propose a coupling-process model (DyCPM) to capture the dynamic mechanisms of them. The model not only generates low-dimensional embedding vectors of nodes, but also aggregates the structural information and temporal information of two kinds of edges. In particular, we design a neural network parameterized discrete process to depict the change law of the topology of evolutive edges and a neural network parameterized temporal point process (TPP) to characterize the temporal dynamic rule of interactive edges. More importantly, we propose a coupling mechanism to transfer the information of the two processes through a shared embedding matrix and finally generate an embedding matrix that aggregates the topology information and temporal information of the two kinds of edges for the dynamic link prediction task. We evaluate our model and several baselines on real datasets. The experimental results show that our model can better aggregate the topology information and temporal information of the two kinds of edges according to their properties and outperforms several state-of-the-art baselines in the performance of dynamic link prediction.

The advance in single-cell profiling technologies and the development in computational algorithms provide the opportunity to reconstruct pseudo temporal trajectory with branch point of cellular development. On the other hand, theories such as dynamical network biomarkers (DNB) theory have been recently proposed to characterize the pre-transition state in biological systems. Few studies have validated whether the branch point identified in pseudo time is the critical point in dynamical system. In this study, the dynamical behavior of the branch point on the pseudo trajectory has been investigated. We study the pseudo temporal trajectories reconstructed by Wishbone and diffusion pseudotime analysis (DPT) algorithms, as well as the simulated trajectory. DNB theory is applied to justify the bifurcating event on the pseudo trajectories. Our results demonstrate that the branch point recovered by Wishbone and DPT algorithms is confirmed as a transition state in cell differentiation process by DNB theory. Furthermore, we show that an appropriate DNB group will amplify the comprehensive index of critical event as defined in DNB theory. Our study provides biological insights on pseudo trajectory with branch point in a dynamical view and also indicates that DNB theory may serve as a benchmark to check the validity of branch point.

In reinforcement learning (RL), one of the key components is policy evaluation, which aims to estimate the value function (i.e., expected long-term accumulated reward) of a policy. With a good policy evaluation method, the RL algorithms will estimate the value function more accurately and find a better policy. When the state space is large or continuous \emph{Gradient-based Temporal Difference(GTD)} policy evaluation algorithms with linear function approximation are widely used. Considering that the collection of the evaluation data is both time and reward consuming, a clear understanding of the finite sample performance of the policy evaluation algorithms is very important to reinforcement learning. Under the assumption that data are i.i.d. generated, previous work provided the finite sample analysis of the GTD algorithms with constant step size by converting them into convex-concave saddle point problems. However, it is well-known that, the data are generated from Markov processes rather than i.i.d in RL problems.. In this paper, in the realistic Markov setting, we derive the finite sample bounds for the general convex-concave saddle point problems, and hence for the GTD algorithms. We have the following discussions based on our bounds. (1) With variants of step size, GTD algorithms converge. (2) The convergence rate is determined by the step size, with the mixing time of the Markov process as the coefficient. The faster the Markov processes mix, the faster the convergence. (3) We explain that the experience replay trick is effective by improving the mixing property of the Markov process. To the best of our knowledge, our analysis is the first to provide finite sample bounds for the GTD algorithms in Markov setting.

BACKGROUND: Recombination plays an important role in the maintenance of genetic diversity in many types of organisms, especially diploid eukaryotes. Recombination can be studied and used to map diseases. However, recombination adds a great deal of complexity to the genetic information. This renders estimation of evolutionary parameters more difficult. After the coalescent process was formulated, models capable of describing recombination using graphs, such as ancestral recombination graphs (ARG) were also developed. There are two typical models based on which to simulate ARG: back-in-time model such as ms and spatial model including Wiuf&Hein's, SMC, SMC', and MaCS. RESULTS: In this study, a new method of modeling coalescence with recombination, Spatial Coalescent simulator (SC), was developed, which considerably improved the algorithm described by Wiuf and Hein. The present algorithm constructs ARG spatially along the sequence, but it does not produce any redundant branches which are inevitable in Wiuf and Hein's algorithm. Interestingly, the distribution of ARG generated by the present new algorithm is identical to that generated by a typical back-in-time model adopted by ms, an algorithm commonly used to model coalescence. It is here demonstrated that the existing approximate methods such as the sequentially Markov coalescent (SMC), a related method called SMC', and Markovian coalescent simulator (MaCS) can be viewed as special cases of the present method. Using simulation analysis, the time to the most common ancestor (TMRCA) in the local trees of ARGs generated by the present algorithm was found to be closer to that produced by ms than time produced by MaCS. Sample-consistent ARGs can be generated using the present method. This may significantly reduce the computational burden. CONCLUSION: In summary, the present method and algorithm may facilitate the estimation and description of recombination in population genomics and evolutionary biology.

This paper proposes a new method for computing page importance, referred to as BrowseRank. The conventional approach to compute page importance is to exploit the link graph of the web and to build a model based on that graph. For instance, PageRank is such an algorithm, which employs a discrete-time Markov process as the model. Unfortunately, the link graph might be incomplete and inaccurate with respect to data for determining page importance, because links can be easily added and deleted by web content creators. In this paper, we propose computing page importance by using a 'user browsing graph' created from user behavior data. In this graph, vertices represent pages and directed edges represent transitions between pages in the users' web browsing history. Furthermore, the lengths of staying time spent on the pages by users are also included. The user browsing graph is more reliable than the link graph for inferring page importance. This paper further proposes using the continuous-time Markov process on the user browsing graph as a model and computing the stationary probability distribution of the process as page importance. An efficient algorithm for this computation has also been devised. In this way, we can leverage hundreds of millions of users' implicit voting on page importance. Experimental results show that BrowseRank indeed outperforms the baseline methods such as PageRank and TrustRank in several tasks.

This paper is concerned with rank aggregation, the task of combining the ranking results of individual rankers at meta-search. Previously, rank aggregation was performed mainly by means of unsupervised learning. To further enhance ranking accuracies, we propose employing supervised learning to perform the task, using labeled data. We refer to the approach as Supervised Rank Aggregation. We set up a general framework for conducting Supervised Rank Aggregation, in which learning is formalized an optimization which minimizes disagreements between ranking results and the labeled data. As case study, we focus on Markov Chain based rank aggregation in this paper. The optimization for Markov Chain based methods is not a convex optimization problem, however, and thus is hard to solve. We prove that we can transform the optimization problem into that of Semidefinite Programming and solve it efficiently. Experimental results on meta-searches show that Supervised Rank Aggregation can significantly outperform existing unsupervised methods.