A Probabilistic Re-Intepretation of Confidence Scores in Multi-Exit Models

In this paper, we propose a new approach to train a deep neural network with multiple intermediate auxiliary classifiers, branching from it. These ‘multi-exits’ models can be used to reduce the inference time by performing early exit on the intermediate branches, if the confidence of the prediction is higher than a threshold. They rely on the assumption that not all the samples require the same amount of processing to yield a good prediction. In this paper, we propose a way to train jointly all the branches of a multi-exit model without hyper-parameters, by weighting the predictions from each branch with a trained confidence score. Each confidence score is an approximation of the real one produced by the branch, and it is calculated and regularized while training the rest of the model. We evaluate our proposal on a set of image classification benchmarks, using different neural models and early-exit stopping criteria.     

Network Bending: Expressive Manipulation of Generative Models in Multiple Domains

This paper presents the network bending framework, a new approach for manipulating and interacting with deep generative models. We present a comprehensive set of deterministic transformations that can be inserted as distinct layers into the computational graph of a trained generative neural network and applied during inference. In addition, we present a novel algorithm for analysing the deep generative model and clustering features based on their spatial activation maps. This allows features to be grouped together based on spatial similarity in an unsupervised fashion. This results in the meaningful manipulation of sets of features that correspond to the generation of a broad array of semantically significant features of the generated results. We outline this framework, demonstrating our results on deep generative models for both image and audio domains. We show how it allows for the direct manipulation of semantically meaningful aspects of the generative process as well as allowing for a broad range of expressive outcomes.     

Permutation Entropy-Based Interpretability of Convolutional Neural Network Models for Interictal EEG Discrimination of Subjects with Epileptic Seizures vs. Psychogenic Non-Epileptic Seizures

The differential diagnosis of epileptic seizures (ES) and psychogenic non-epileptic seizures (PNES) may be difficult, due to the lack of distinctive clinical features. The interictal electroencephalographic (EEG) signal may also be normal in patients with ES. Innovative diagnostic tools that exploit non-linear EEG analysis and deep learning (DL) could provide important support to physicians for clinical diagnosis. In this work, 18 patients with new-onset ES (12 males, 6 females) and 18 patients with video-recorded PNES (2 males, 16 females) with normal interictal EEG at visual inspection were enrolled. None of them was taking psychotropic drugs. A convolutional neural network (CNN) scheme using DL classification was designed to classify the two categories of subjects (ES vs. PNES). The proposed architecture performs an EEG time-frequency transformation and a classification step with a CNN. The CNN was able to classify the EEG recordings of subjects with ES vs. subjects with PNES with 94.4% accuracy. CNN provided high performance in the assigned binary classification when compared to standard learning algorithms (multi-layer perceptron, support vector machine, linear discriminant analysis and quadratic discriminant analysis). In order to interpret how the CNN achieved this performance, information theoretical analysis was carried out. Specifically, the permutation entropy (PE) of the feature maps was evaluated and compared in the two classes. The achieved results, although preliminary, encourage the use of these innovative techniques to support neurologists in early diagnoses.     

Electric Double Layer in Water-Organic Mixed Solvents: Titania in 50% Ethylene Glycol

Ethylene glycol (EG) and its mixtures with water are popular components of nanofluids used as heat transfer fluids. The stability of nanofluids against coagulation is correlated with their zeta potential. The electrophoretic mobility of titania nanoparticles in 50-50 w/w EG was studied as a function of the concentration of various solutes. HCl, NaOH, SDS and CTMABr at concentrations up to 0.01 M are strong electrolytes in 50% EG, that is, the conductance of their solutions is proportional to the concentration. HCl, NaOH and CTMABr were very efficient in inducing a high zeta potential for titania in 50% EG. NaOH induced a negative zeta potential in excess of 70 mV, and HCl and CTMABr induced a positive zeta potential in excess of 70 mV at concentrations below 10⁻⁴ M. Apparently, HCl, NaOH and CTMABr are also more efficient than SDS in terms of nanofluid stabilization against coagulation. An overdose of base (>1 mM) results in depression of the negative zeta potential. This result may be due to the specific adsorption of sodium on titania from 50% EG.     

Effect of Natural Deep Eutectic Solvents on trans-Resveratrol Photo-Chemical Induced Isomerization and 2,4,6-Trihydroxyphenanthrene Electro-Cyclic Formation

trans-Resveratrol is a natural bioactive compound with well-recognized health promoting effects. When exposed to UV light, this compound can undergo a photochemically induced trans/cis isomerization and a 6π electrochemical cyclization with the subsequent formation of 2,4,6-trihydroxyphenanthrene (THP). THP is a potentially harmful compound which can exert genotoxic effects. In this work we improved the chromatographic separation and determination of the two resveratrol isomers and of THP by using a non-commercial pentafluorophenyl stationary phase. We assessed the effect of natural deep eutectic solvents (NaDES) as possible photo-protective agents by evaluating cis-resveratrol isomer and THP formation under different UV-light exposure conditions with the aim of enhancing resveratrol photostability and inhibiting THP production. Our results demonstrate a marked photoprotective effect exerted by glycerol-containing NaDES, and in particular by proline/glycerol NaDES, which exerts a strong inhibitory effect on the photochemical isomerization of resveratrol and significantly limits the formation of the toxic derivative THP. Considering the presence of resveratrol in various commercial products, these results are of note in view of the potential genotoxic risk associated with its photochemical degradation products and in view of the need for the development of green, eco-sustainable and biocompatible resveratrol photo-stable formulations.     

Deep Hierarchical Ensemble Model for Suicide Detection on Imbalanced Social Media Data

As a serious worldwide problem, suicide often causes huge and irreversible losses to families and society. Therefore, it is necessary to detect and help individuals with suicidal ideation in time. In recent years, the prosperous development of social media has provided new perspectives on suicide detection, but related research still faces some difficulties, such as data imbalance and expression implicitness. In this paper, we propose a Deep Hierarchical Ensemble model for Suicide Detection (DHE-SD) based on a hierarchical ensemble strategy, and construct a dataset based on Sina Weibo, which contains more than 550 thousand posts from 4521 users. To verify the effectiveness of the model, we also conduct experiments on a public Weibo dataset containing 7329 users’ posts. The proposed model achieves the best performance on both the constructed dataset and the public dataset. In addition, in order to make the model applicable to a wider population, we use the proposed sentence-level mask mechanism to delete user posts with strong suicidal ideation. Experiments show that the proposed model can still effectively identify social media users with suicidal ideation even when the performance of the baseline models decrease significantly.     

从科研中提炼出的精细化工专业实验:新型疏水缔合聚合物的制备、表征及性能测定*

结合教师科研项目,推荐一个精细化工专业实验:疏水缔合聚合物的制备、表征及性能测定。通过该实验可以使学生掌握自由基共聚法制备聚合物的基本原理,掌握一些大型仪器的基本操作、聚合物常用的表征手段及结构分析方法,理解疏水缔合聚合物结构与性能之间深层次的理论联系。该实验项目来源于科研,服务于实践,集聚合反应、结构表征、大分子溶液理化性能测试于一体,具有综合性、前沿性、研究性等特点。本实验的开设有利于提高精细化工方向学生的专业实验水平,提升学生的综合探究素质、创新意识及动手能力。