Empirical Analysis of Multi-T ask Learning for Reducing Model Bias in T oxic Comment Detection Ameya V aidya, 1 Feng Mai, 2 Y ue Ning 3 1 Bridgewater-Raritan Regional High School 2 School of Business, Stevens Institute of Technology 3 Department of Computer Science, Stevens Institute of Technology ameyav993@gmail.com, Abstract With the recent rise of toxicity in online conversations on social media platforms, using modern machine learning algorithms for toxic comment detection has become a central focus of many online applications. Researchers and companies have developed a variety of shallow and deep learning models to identify toxicity in online conversations, reviews, or comments with mixed successes. However, these existing approaches have learned to incorrectly associate nontoxic comments that have certain trigger-words (e.g. In this paper, we evaluate dozens of state-of-the-art models with the specific focus of reducing model bias towards these commonly-attacked identity groups. We propose a multi-task learning model with an attention layer that jointly learns to predict the toxicity of a comment as well as the identities present in the comments in order to reduce this bias. We then compare our model to an array of shallow and deep-learning models using metrics designed especially to test for unintended model bias within these identity groups. Introduction The identification of potential toxicity within online conversations has always been a significant task for current platform providers. Toxic comments have the unfortunate effect of causing users to leave a discussion or give up sharing their perspective and can give a bad reputation to platforms where these discussions take place. Twitter's CEO reaffirmed that Twitter is still being overrun by spam, abuse, and misinformation. Current research involves investigating common challenges in toxic comment classification (van Aken et al. 2018), identifying subtle forms of toxicity (Noever 2018), detecting early signs of toxicity (Zhang et al. 2018), and analysing sarcasm within conversations (Ghosh, Fabbri, and Muresan 2018).

Recently, pretrained language models (e.g., BERT) have achieved great success on many downstream natural language understanding tasks and exhibit a certain level of commonsense reasoning ability. However, their performance on commonsense tasks is still far from that of humans. As a preliminary attempt, we propose a simple yet effective method to teach pretrained models with commonsense reasoning by leveraging the structured knowledge in ConceptNet, the largest commonsense knowledge base (KB). Specifically, the structured knowledge in KB allows us to construct various logical forms, and then generate multiple-choice questions requiring commonsense logical reasoning. Experimental results demonstrate that, when refined on these training examples, the pretrained models consistently improve their performance on tasks that require commonsense reasoning, especially in the few-shot learning setting. Besides, we also perform analysis to understand which logical relations are more relevant to commonsense reasoning.

We build a bridge between neural network-based machine learning and graph-based natural language processing and introduce a unified approach to keyphrase, summary and relation extraction by aggregating dependency graphs from links provided by a deep-learning based dependency parser. We reorganize dependency graphs to focus on the most relevant content elements of a sentence, integrate sentence identifiers as graph nodes and after ranking the graph, we extract our keyphrases and summaries from its largest strongly-connected component. We take advantage of the implicit structural information that dependency links bring to extract subject-verb-object, is-a and part-of relations. We put it all together into a proof-of-concept dialog engine that specializes the text graph with respect to a query and reveals interactively the document's most relevant content elements. The open-source code of the integrated system is available at https:// github.com/ptarau/DeepRank .

With the rapid increase in cloud computing, concerns surrounding data privacy, security, and confidentiality also have been increased significantly. Not only cloud providers are susceptible to internal and external hacks, but also in some scenarios, data owners cannot outsource the computation due to privacy laws such as GDPR, HIPAA, or CCPA. Fully Homomorphic Encryption (FHE) is a groundbreaking invention in cryptography that, unlike traditional cryptosystems, enables computation on encrypted data without ever decrypting it. However, the most critical obstacle in deploying FHE at large-scale is the enormous computation overhead. In this paper, we present HEAX, a novel hardware architecture for FHE that achieves unprecedented performance improvement. HEAX leverages multiple levels of parallelism, ranging from ciphertext-level to fine-grained modular arithmetic level. Our first contribution is a new highly-parallelizable architecture for number-theoretic transform (NTT) which can be of independent interest as NTT is frequently used in many lattice-based cryptography systems. Building on top of NTT engine, we design a novel architecture for computation on homomorphically encrypted data. We also introduce several techniques to enable an end-to-end, fully pipelined design as well as reducing on-chip memory consumption. Our implementation on reconfigurable hardware demonstrates 164-268x performance improvement for a wide range of FHE parameters.

Conceptual entanglement is a crucial phenomenon in quantum cognition because it implies that classical probabilities cannot model non--compositional conceptual phenomena. While several psychological experiments have been developed to test conceptual entanglement, this has not been explored in the context of Natural Language Processing. In this paper, we apply the hypothesis that words of a document are traces of the concepts that a person has in mind when writing the document. Therefore, if these concepts are entangled, we should be able to observe traces of their entanglement in the documents. In particular, we test conceptual entanglement by contrasting language simulations with results obtained from a text corpus. Our analysis indicates that conceptual entanglement is strongly linked to the way in which language is structured. We discuss the implications of this finding in the context of conceptual modeling and of Natural Language Processing.

One of the most studied problems in machine learning is finding reasonable constraints that guarantee the generalization of a learning algorithm. These constraints are usually expressed as some simplicity assumptions on the target. For instance, in the V apnik-Chervonenkis (VC) theory the space of possible hypotheses is considered to have a limited VC dimension and in kernel methods there are assumptions on the spectrum of the operator in the Hilbert space. One way to formulate the simplicity assumption is via information theoretic concepts. In this paper, the constraint on the entropy H ( X) of the input variable X is studied as a simplicity assumption. It is proven that the sample complexity to achieve an null -δ Probably Approximately Correct (P AC) hypothesis is bounded by 2 2 H ( X) / null log 1 δ null 2 which is sharp up to the 1 null 2 factor. Morever, it is shown that if a feature learning process is employed to learn the compressed representation from the dataset, this bound no longer exists. These findings have important implications on the Information Bottleneck (IB) theory which had been utilized to explain the generalization power of Deep Neural Networks (DNNs), but its applicability for this purpose is currently under debate by researchers. In particular, this is a rigorous proof for the previous heuristic that compressed representations are expnentially easier to be learned. However, our analysis pinpoints two factors preventing the IB, in its current form, to be applicable in studying neural networks. Firstly, the exponential dependence of sample complexity on 1 / null, which can lead to a dramatic e ff ect on the bounds in practical applications when null is small. Secondly, our analysis reveals that arguments based on input compression are inherently insu fficient to explain generalization of methods like DNNs in which the features are also learned using available data. Keywords: Compressed Representation; Generalization Bound; Information Bottleneck. 1. Introduction The main objective of learning is to develop algorithms which can learn general patterns by using a finite number of samples drawn from a target distribution. The "no free lunch" theorem states that if there is no constraint on the distribution, it is impossible to say anything about the samples not seen in the training set (Wolpert, 1996b).

We propose a planning and perception mechanism for a robot (agent), that can only observe the underlying environment partially, in order to solve an image classification problem. A three-layer architecture is suggested that consists of a meta-layer that decides the intermediate goals, an action-layer that selects local actions as the agent navigates towards a goal, and a classification-layer that evaluates the reward and makes a prediction. We design and implement these layers using deep reinforcement learning. A generalized policy gradient algorithm is utilized to learn the parameters of these layers to maximize the expected reward. Our proposed methodology is tested on the MNIST dataset of handwritten digits, which provides us with a level of explainability while interpreting the agent's intermediate goals and course of action.

Most advances in single image de-raining meet a key challenge, which is removing rain streaks with different scales and shapes while preserving image details. Existing single image de-raining approaches treat rain-streak removal as a process of pixel-wise regression directly. However, they are lacking in mining the balance between over-de-raining (e.g. removing texture details in rain-free regions) and under-de-raining (e.g. leaving rain streaks). In this paper, we firstly propose a coarse-to-fine network called Gradual Network (GraNet) consisting of coarse stage and fine stage for delving into single image de-raining with different granularities. Specifically, to reveal coarse-grained rain-streak characteristics (e.g. long and thick rain streaks/raindrops), we propose a coarse stage by utilizing local-global spatial dependencies via a local-global subnetwork composed of region-aware blocks. Taking the residual result (the coarse de-rained result) between the rainy image sample (i.e. the input data) and the output of coarse stage (i.e. the learnt rain mask) as input, the fine stage continues to de-rain by removing the fine-grained rain streaks (e.g. light rain streaks and water mist) to get a rain-free and well-reconstructed output image via a unified contextual merging sub-network with dense blocks and a merging block. Solid and comprehensive experiments on synthetic and real data demonstrate that our GraNet can significantly outperform the state-of-the-art methods by removing rain streaks with various densities, scales and shapes while keeping the image details of rain-free regions well-preserved.

Differentiable Architecture Search (DARTS) has attracted a lot of attention due to its simplicity and small search costs achieved by a continuous relaxation and an approximation of the resulting bi-level optimization problem. However, DARTS does not work robustly for new problems: we identify a wide range of search spaces for which DARTS yields degenerate architectures with very poor test performance. We study this failure mode and show that, while DARTS successfully minimizes validation loss, the found solutions generalize poorly when they coincide with high validation loss curvature in the space of architectures. We show that by adding one of various types of regularization we can robustify DARTS to find solutions with smaller Hessian spectrum and with better generalization properties. Based on these observations we propose several simple variations of DARTS that perform substantially more robustly in practice. Our observations are robust across five search spaces on three image classification tasks and also hold for the very different domains of disparity estimation (a dense regression task) and language modelling. We provide our implementation and scripts to facilitate reproducibility.

Bike Sharing Systems (BSSs) have been adopted in many major cities of the world due to traffic congestion and carbon emissions. Although there have been approaches to exploiting either bike trailers via crowdsourcing or carrier vehicles to reposition bikes in the "right" stations in the "right" time, they do not jointly consider the usage of both bike trailers and carrier vehicles. In this paper, we aim to take advantage of both bike trailers and carrier vehicles to reduce the loss of demand with regard to the crowdsourcing of bike trailers and the fuel cost of carrier vehicles. In the experiment, we exhibit that our approach outperforms baselines in several datasets from bike sharing companies. Introduction Bike sharing systems (BSSs) typically have a set of base stations that are strategically placed throughout a city and each station has a fixed number of docks, e.g., Capital Bike-share 1, Bluebikes 2, Mobike 3, BIXI 4, etc. At the beginning of the day, each station is stocked with a predetermined number of bikes. Customers can pick and drop bikes from any station and are charged depending on the hiring duration (Tsai, Chen, and Hong 2019; Hulot, Aloise, and Jena 2018; Lowalekar et al. 2017; Vulcano, van Ryzin, and Ratliff 2012; Schuijbroek, Hampshire, and van Hoeve 2017). Due to the individualistic and uncoordinated movements of customers, there is often starvation (empty base stations precluding bike pickup) or congestion (full base stations precluding bike return) of bikes at certain stations, which results in a significant loss of customer demand (Shu et al. 2013; Chen, Liu, and Liu 2018). To address this problem, a variety of systems (Ghosh et al. 2017; Lowalekar et al. 2017) employ the idea of repositioning idle bikes with the help of carrier vehicles during the day, by taking into account the movement of bikes by customers (Tsai, Chen, and Hong 2019; Pfrommer et al. 2014; Ghosh and V arakantham 2017).