This paper shows that standard assessment methodology for style transfer has several significant problems. First, the standard metrics for style accuracy and semantics preservation vary significantly on different re-runs. Therefore one has to report error margins for the obtained results. Second, starting with certain values of bilingual evaluation understudy (BLEU) between input and output and accuracy of the sentiment transfer the optimization of these two standard metrics diverge from the intuitive goal of the style transfer task. Finally, due to the nature of the task itself, there is a specific dependence between these two metrics that could be easily manipulated. Under these circumstances, we suggest taking BLEU between input and human-written reformulations into consideration for benchmarks. We also propose three new architectures that outperform state of the art in terms of this metric.

With the advent of deep generative models in computational chemistry, in silico anticancer drug design has undergone an unprecedented transformation. While state-of-the-art deep learning approaches have shown potential in generating compounds with desired chemical properties, they entirely overlook the genetic profile and properties of the target disease. In the case of cancer, this is problematic since it is a highly genetic disease in which the biomolecular profile of target cells determines the response to therapy. Here, we introduce the first deep generative model capable of generating anticancer compounds given a target biomolecular profile. Using a reinforcement learning framework, the transcriptomic profile of cancer cells is used as a context in which anticancer molecules are generated and optimized to obtain effective compounds for the given profile. Our molecule generator combines two pretrained variational autoencoders (VAEs) and a multimodal efficacy predictor - the first VAE generates transcriptomic profiles while the second conditional VAE generates novel molecular structures conditioned on the given transcriptomic profile. The efficacy predictor is used to optimize the generated molecules through a reward determined by the predicted IC50 drug sensitivity for the generated molecule and the target profile. We demonstrate how the molecule generation can be biased towards compounds with high inhibitory effect against individual cell lines or specific cancer sites. We verify our approach by investigating candidate drugs generated against specific cancer types and investigate their structural similarity to existing compounds with known efficacy against these cancer types. We envision our approach to transform in silico anticancer drug design by increasing success rates in lead compound discovery via leveraging the biomolecular characteristics of the disease.

Training Optimus Prime, M.D.: Generating Medical Certification Items by Fine-Tuning OpenAI's gpt2 Transformer Model Matthias von Davier August 21st, 2019 Abstract Objective: Showcasing Artificial Intelligence, in particular deep neural networks, for language modeling aimed at automated generation of medical education test items. Materials and Methods: OpenAI's gpt2 transformer language model was retrained using PubMed's open access text mining database. The retraining was done using toolkits based on tensorflow-gpu available on GitHub, using a workstation equipped with two GPUs. Results: In comparison to a study that used character based recurrent neural networks trained on open access items, the retrained transformer architecture allows generating higher quality text that can be used as draft input for medical education assessment material. In addition, prompted text generation can be used for production of distractors suitable for multiple choice items used in certification exams. Discussion: The current state of neural network based language models can be used to develop tools in supprt of authoring medical education exams using retrained models on the basis of corpora consisting of general medical text collections. Conclusion: Future experiments with more recent transformer models (such as Grover, TransformerXL) using existing medical certification exam item pools is expected to further improve results and facilitate the development of assessment materials. Objective The aim of this article is to provide evidence on the current state of automated item generation (AIG) using deep neural networks (DNNs). Based on earlier work, a first paper that tackled this issue used character-based Address for correspondence: mvondavier@nbme.org: Time flies in the domain of DNNs used for language modeling, indeed: The day this paper was submitted, on August 13th, 2019, to internal review, NVIDIA published yet another, larger language model of the transformer used in this paper. The MegratronLM (apart from taking a bite out of the pun in this article's title) is currently the largest language model based on the transformer architecture [3]. This latest neural network language model has 8 billions of parameters, which is incomprehensible compared to the type of neural networks we used only two decades ago. At that time, in winter semester 1999-2000, I taught classes about artificial Neural Networks (NNs, e.g. Back then, Artificial Intelligence (AI) already entered what was referred to as AI winter, as most network sizes were limited to rather small architectures unless supercomputers were employed.

We propose a novel framework, learning to transfer learn (L2TL), to improve transfer learning on a target dataset by judicious extraction of information from a source dataset. Our framework considers joint optimization of strongly-shared weights between models of source and target tasks, and employs adaptive weights for scaling of constituent loss terms. The adaptation of the weights is done using a reinforcement learning (RL)-based policy model, which is guided based on a performance metric on the target validation set. We demonstrate state-of-the-art performance of L2TL given fixed models, consistently outperforming fine-tuning baselines on various datasets. In addition, in the regimes of small-scale target datasets and significant label mismatch between source and target datasets, L2TL outperforms previous methods by a large margin.

Online purchase decisions in organizations can go through a complex journey with multiple agents involved in the decision making process. Depending on the product being purchased, and the organizational structure, the process may involve employees who first conduct market research, and then influence decision makers who place the online purchase order. In such cases, the online activity trail of a single individual in the organization may only provide partial information for predicting purchases (conversions). To refine conversion prediction for business-to-business (B2B) products using online activity trails, we introduce the notion of relevant users in an organization with respect to a given B2B advertiser, and leverage the collective activity trails of such relevant users to predict conversions. In particular, our notion of relevant users is tied to a seed list of relevant activities for a B2B advertiser, and we propose a method using distributed activity representations to build such a seed list. Experiments using data from Yahoo Gemini demonstrate that the proposed methods can improve conversion prediction AUC by 8.8%, and provide an interpretable advertiser specific list of activities useful for B2B ad targeting.

With the huge amount of information inflowing the World Wide Web every second, it becomes more difficult and more difficult to retrieve information from the Web. This explains why the existence of a search engine is as important as the existence of the web itself. Since the appearance of the web, there has been a fundamental talk in the web research communit y to develop the rapid, effective, and precise search engines. This paper will be chiefly discussing about the most common search engine nowadays which is Google. The mathematical theory behind the Google search engine is the PageRank algorithm, which was presented by Sergey Brin and Lawrence Page [1].

Preprint submitted to RGN Publications on 21 /5/2018 Abstract The credit cards' fraud transactions detection is the important problem in machine learning field. To detect the credit cards' fraud transactions help reduce the significant loss of the credit cards' holders and the banks. To detect the credit cards' fraud transactions, data scientists normally employ the un - supervised learning techniques and supervised learning technique. In this paper, we employ the graph p - Laplacian based semi - supervised learning methods combi ned with the under - sampling technique such as Cluster Centroids to solve the credit cards' fraud transactions detection problem. Experimental results show that that the graph p - Laplacian semi - supervised learning method s outper form the current state of art graph Laplacian based semi - supervised learning method ( p 2). 2010 AMS Classi fi cation: 05C85 Keywords and phrases: graph p - Laplacian, credit card, fraud detection, semi - supervised learning Article type: Research article 1 Introduction While purchasing online, the transactions can be done by using credit cards that are issued by the bank.

It is well known in astronomy that propagating non-Gaussian prediction uncertainty in photometric redshift estimates is key to reducing bias in downstream cosmological analyses. Similarly, likelihood-free inference approaches, which are beginning to emerge as a tool for cosmological analysis, require the full uncertainty landscape of the parameters of interest given observed data. However, most machine learning (ML) based methods with open-source software target point prediction or classification, and hence fall short in quantifying uncertainty in complex regression and parameter inference settings such as the applications mentioned above. As an alternative to methods that focus on predicting the response (or parameters) $\mathbf{y}$ from features $\mathbf{x}$, we provide nonparametric conditional density estimation (CDE) tools for approximating and validating the entire probability density $\mathrm{p}(\mathbf{y} \mid \mathbf{x})$ given training data for $\mathbf{x}$ and $\mathbf{y}$. As there is no one-size-fits-all CDE method, the goal of this work is to provide a comprehensive range of statistical tools and open-source software for nonparametric CDE and method assessment which can accommodate different types of settings and which in addition can easily be fit to the problem at hand. Specifically, we introduce CDE software packages in $\texttt{Python}$ and $\texttt{R}$ based on four ML prediction methods adapted and optimized for CDE: $\texttt{NNKCDE}$, $\texttt{RFCDE}$, $\texttt{FlexCode}$, and $\texttt{DeepCDE}$. Furthermore, we present the $\texttt{cdetools}$ package, which includes functions for computing a CDE loss function for model selection and tuning of parameters, together with diagnostics functions. We provide sample code in $\texttt{Python}$ and $\texttt{R}$ as well as examples of applications to photometric redshift estimation and likelihood-free cosmology via CDE.

Network Embedding is the task of learning continuous node representations for networks, which has been shown effective in a variety of tasks such as link prediction and node classification. Most of existing works aim to preserve different network structures and properties in low-dimensional embedding vectors, while neglecting the existence of noisy information in many real-world networks and the overfitting issue in the embedding learning process. Most recently, generative adversarial networks (GANs) based regularization methods are exploited to regularize embedding learning process, which can encourage a global smoothness of embedding vectors. These methods have very complicated architecture and suffer from the well-recognized non-convergence problem of GANs. In this paper, we aim to introduce a more succinct and effective local regularization method, namely adversarial training, to network embedding so as to achieve model robustness and better generalization performance. Firstly, the adversarial training method is applied by defining adversarial perturbations in the embedding space with an adaptive $L_2$ norm constraint that depends on the connectivity pattern of node pairs. Though effective as a regularizer, it suffers from the interpretability issue which may hinder its application in certain real-world scenarios. To improve this strategy, we further propose an interpretable adversarial training method by enforcing the reconstruction of the adversarial examples in the discrete graph domain. These two regularization methods can be applied to many existing embedding models, and we take DeepWalk as the base model for illustration in the paper. Empirical evaluations in both link prediction and node classification demonstrate the effectiveness of the proposed methods.

We present a queuing model of parking dynamics and a model-based prediction method to provide real-time probabilistic forecasts of future parking occupancy. The queuing model has a non-homogeneous arrival rate and time-varying service time distribution. All statistical assumptions of the model are verified using data from 29 truck parking locations, each with between 55 and 299 parking spots. For each location and each spot the data specifies the arrival and departure times of a truck, for 16 months of operation. The modeling framework presented in this paper provides empirical support for queuing models adopted in many theoretical studies and policy designs. We discuss how our framework can be used to study parking problems in different environments. Based on the queuing model, we propose two prediction methods, a microscopic method and a macroscopic method, that provide a real-time probabilistic forecast of parking occupancy for an arbitrary forecast horizon. These model-based methods convert a probabilistic forecast problem into a parameter estimation problem that can be tackled using classical estimation methods such as regressions or pure machine learning algorithms. We characterize a lower bound for an arbitrary real-time prediction algorithm. We evaluate the performance of these methods using the truck data comparing the outcomes of their implementations with other model-based and model-free methods proposed in the literature.