Paper Group ANR 1015
An Information-Theoretic Perspective on the Relationship Between Fairness and Accuracy. Acoustic-to-Word Models with Conversational Context Information. Mimicking Human Process: Text Representation via Latent Semantic Clustering for Classification. An Attention-Guided Deep Regression Model for Landmark Detection in Cephalograms. Generalized Transfo …
An Information-Theoretic Perspective on the Relationship Between Fairness and Accuracy
| Title | An Information-Theoretic Perspective on the Relationship Between Fairness and Accuracy |
| Authors | Sanghamitra Dutta, Dennis Wei, Hazar Yueksel, Pin-Yu Chen, Sijia Liu, Kush R. Varshney |
| Abstract | Our goal is to understand the so-called trade-off between fairness and accuracy. In this work, using a tool from information theory called Chernoff information, we derive fundamental limits on this relationship that explain why the accuracy on a given dataset often decreases as fairness increases. Novel to this work, we examine the problem of fair classification through the lens of a mismatched hypothesis testing problem, i.e., where we are trying to find a classifier that distinguishes between two “ideal” distributions but instead we are given two mismatched distributions that are biased. Based on this perspective, we contend that measuring accuracy with respect to the given (possibly biased) dataset is a problematic measure of performance. Instead one should also consider accuracy with respect to an ideal dataset that is unbiased. We formulate an optimization to find such ideal distributions and show that the optimization is feasible. Lastly, when the Chernoff information for one group is strictly less than another in the given dataset, we derive the information-theoretic criterion under which collection of more features can actually improve the Chernoff information and achieve fairness without compromising accuracy on the available data. |
| Tasks | |
| Published | 2019-10-17 |
| URL | https://arxiv.org/abs/1910.07870v1 |
| https://arxiv.org/pdf/1910.07870v1.pdf | |
| PWC | https://paperswithcode.com/paper/an-information-theoretic-perspective-on-the |
| Repo | |
| Framework | |
Acoustic-to-Word Models with Conversational Context Information
| Title | Acoustic-to-Word Models with Conversational Context Information |
| Authors | Suyoun Kim, Florian Metze |
| Abstract | Conversational context information, higher-level knowledge that spans across sentences, can help to recognize a long conversation. However, existing speech recognition models are typically built at a sentence level, and thus it may not capture important conversational context information. The recent progress in end-to-end speech recognition enables integrating context with other available information (e.g., acoustic, linguistic resources) and directly recognizing words from speech. In this work, we present a direct acoustic-to-word, end-to-end speech recognition model capable of utilizing the conversational context to better process long conversations. We evaluate our proposed approach on the Switchboard conversational speech corpus and show that our system outperforms a standard end-to-end speech recognition system. |
| Tasks | End-To-End Speech Recognition, Speech Recognition |
| Published | 2019-05-21 |
| URL | https://arxiv.org/abs/1905.08796v1 |
| https://arxiv.org/pdf/1905.08796v1.pdf | |
| PWC | https://paperswithcode.com/paper/acoustic-to-word-models-with-conversational |
| Repo | |
| Framework | |
Mimicking Human Process: Text Representation via Latent Semantic Clustering for Classification
| Title | Mimicking Human Process: Text Representation via Latent Semantic Clustering for Classification |
| Authors | Xiaoye Tan, Rui Yan, Chongyang Tao, Mingrui Wu |
| Abstract | Considering that words with different characteristic in the text have different importance for classification, grouping them together separately can strengthen the semantic expression of each part. Thus we propose a new text representation scheme by clustering words according to their latent semantics and composing them together to get a set of cluster vectors, which are then concatenated as the final text representation. Evaluation on five classification benchmarks proves the effectiveness of our method. We further conduct visualization analysis showing statistical clustering results and verifying the validity of our motivation. |
| Tasks | |
| Published | 2019-06-18 |
| URL | https://arxiv.org/abs/1906.07525v1 |
| https://arxiv.org/pdf/1906.07525v1.pdf | |
| PWC | https://paperswithcode.com/paper/mimicking-human-process-text-representation |
| Repo | |
| Framework | |
An Attention-Guided Deep Regression Model for Landmark Detection in Cephalograms
| Title | An Attention-Guided Deep Regression Model for Landmark Detection in Cephalograms |
| Authors | Zhusi Zhong, Jie Li, Zhenxi Zhang, Zhicheng Jiao, Xinbo Gao |
| Abstract | Cephalometric tracing method is usually used in orthodontic diagnosis and treat-ment planning. In this paper, we propose a deep learning based framework to au-tomatically detect anatomical landmarks in cephalometric X-ray images. We train the deep encoder-decoder for landmark detection, and combine global landmark configuration with local high-resolution feature responses. The proposed frame-work is based on 2-stage u-net, regressing the multi-channel heatmaps for land-mark detection. In this framework, we embed attention mechanism with global stage heatmaps, guiding the local stage inferring, to regress the local heatmap patches in a high resolution. Besides, the Expansive Exploration strategy im-proves robustness while inferring, expanding the searching scope without in-creasing model complexity. We have evaluated our framework in the most wide-ly-used public dataset of landmark detection in cephalometric X-ray images. With less computation and manually tuning, our framework achieves state-of-the-art results. |
| Tasks | |
| Published | 2019-06-17 |
| URL | https://arxiv.org/abs/1906.07549v1 |
| https://arxiv.org/pdf/1906.07549v1.pdf | |
| PWC | https://paperswithcode.com/paper/an-attention-guided-deep-regression-model-for |
| Repo | |
| Framework | |
Generalized Transformation-based Gradient
| Title | Generalized Transformation-based Gradient |
| Authors | Anbang Wu, Shuangxi Chen, Chunming Wu |
| Abstract | The reparameterization trick has become one of the most useful tools in the field of variational inference. However, the reparameterization trick is based on the standardization transformation which restricts the scope of application of this method to distributions that have tractable inverse cumulative distribution functions or are expressible as deterministic transformations of such distributions. In this paper, we generalized the reparameterization trick by allowing a general transformation. We discover that the proposed model is a special case of control variate indicating that the proposed model can combine the advantages of CV and generalized reparameterization. |
| Tasks | |
| Published | 2019-11-06 |
| URL | https://arxiv.org/abs/1911.02681v3 |
| https://arxiv.org/pdf/1911.02681v3.pdf | |
| PWC | https://paperswithcode.com/paper/generalized-transformation-based-gradient |
| Repo | |
| Framework | |
A generalized intelligent quality-based approach for fusing multi-source information
| Title | A generalized intelligent quality-based approach for fusing multi-source information |
| Authors | Fuyuan Xiao |
| Abstract | In this paper, we propose a generalized intelligent quality-based approach for fusing multi-source information. The goal of the proposed approach intends to fuse the multi-complex-valued distribution information while maintaining a high quality of the fused result by considering the usage of credible information sources. |
| Tasks | |
| Published | 2019-10-11 |
| URL | https://arxiv.org/abs/1910.06150v1 |
| https://arxiv.org/pdf/1910.06150v1.pdf | |
| PWC | https://paperswithcode.com/paper/a-generalized-intelligent-quality-based |
| Repo | |
| Framework | |
Graph Neural Lasso for Dynamic Network Regression
| Title | Graph Neural Lasso for Dynamic Network Regression |
| Authors | Yixin Chen, Lin Meng, Jiawei Zhang |
| Abstract | In this paper, we will study the dynamic network regression problem, which focuses on inferring both individual entities’ changing attribute values and the dynamic relationships among the entities in the network data simultaneously. To resolve the problem, a novel graph neural network, namely graph neural lasso (GNL), will be proposed in this paper. To model the real-time changes of nodes in the network, GNL extends gated diffusive unit (GDU) to the regression scenario and uses it as the basic neuron unit. GNL can effectively model the dynamic relationships among the nodes based on an attention mechanism. |
| Tasks | |
| Published | 2019-07-25 |
| URL | https://arxiv.org/abs/1907.11114v1 |
| https://arxiv.org/pdf/1907.11114v1.pdf | |
| PWC | https://paperswithcode.com/paper/graph-neural-lasso-for-dynamic-network |
| Repo | |
| Framework | |
Active Search for Nearest Neighbors
| Title | Active Search for Nearest Neighbors |
| Authors | Hayoung Um, Heeyoul Choi |
| Abstract | In pattern recognition or machine learning, it is a very fundamental task to find nearest neighbors of a given point. All the methods for the task work basically by comparing the given point to all the points in the data set. That is why the computational cost increases with the number of data points. However, the human visual system seems to work in a different way. When the human visual system tries to find the neighbors of one point on a map, it directly focuses on the area around the point and actively searches the neighbors by looking or zooming in and out around the point. In this paper, we propose an innovative search method for nearest neighbors, which seems very similar to how human visual system works on the task. |
| Tasks | |
| Published | 2019-12-01 |
| URL | https://arxiv.org/abs/1912.00386v2 |
| https://arxiv.org/pdf/1912.00386v2.pdf | |
| PWC | https://paperswithcode.com/paper/active-search-for-nearest-neighbors |
| Repo | |
| Framework | |
Sharp Bounds on the Runtime of the (1+1) EA via Drift Analysis and Analytic Combinatorial Tools
| Title | Sharp Bounds on the Runtime of the (1+1) EA via Drift Analysis and Analytic Combinatorial Tools |
| Authors | Hsien-Kuei Hwang, Carsten Witt |
| Abstract | The expected running time of the classical (1+1) EA on the OneMax benchmark function has recently been determined by Hwang et al. (2018) up to additive errors of $O((\log n)/n)$. The same approach proposed there also leads to a full asymptotic expansion with errors of the form $O(n^{-K}\log n)$ for any $K>0$. This precise result is obtained by matched asymptotics with rigorous error analysis (or by solving asymptotically the underlying recurrences via inductive approximation arguments), ideas radically different from well-established techniques for the running time analysis of evolutionary computation such as drift analysis. This paper revisits drift analysis for the (1+1) EA on OneMax and obtains that the expected running time $E(T)$, starting from $\lceil n/2\rceil$ one-bits, is determined by the sum of inverse drifts up to logarithmic error terms, more precisely $$\sum_{k=1}^{\lfloor n/2\rfloor}\frac{1}{\Delta(k)} - c_1\log n \le E(T) \le \sum_{k=1}^{\lfloor n/2\rfloor}\frac{1}{\Delta(k)} - c_2\log n,$$ where $\Delta(k)$ is the drift (expected increase of the number of one-bits from the state of $n-k$ ones) and $c_1,c_2 >0$ are explicitly computed constants. This improves the previous asymptotic error known for the sum of inverse drifts from $\tilde{O}(n^{2/3})$ to a logarithmic error and gives for the first time a non-asymptotic error bound. Using standard asymptotic techniques, the difference between $E(T)$ and the sum of inverse drifts is found to be $(e/2)\log n+O(1)$. |
| Tasks | |
| Published | 2019-06-21 |
| URL | https://arxiv.org/abs/1906.09047v2 |
| https://arxiv.org/pdf/1906.09047v2.pdf | |
| PWC | https://paperswithcode.com/paper/sharp-bounds-on-the-runtime-of-the-11-ea-via |
| Repo | |
| Framework | |
TE-ETH: Lower Bounds for QBFs of Bounded Treewidth
| Title | TE-ETH: Lower Bounds for QBFs of Bounded Treewidth |
| Authors | Johannes Klaus Fichte, Markus Hecher, Andreas Pfandler |
| Abstract | The problem of deciding the validity (QSAT) of quantified Boolean formulas (QBF) is a vivid research area in both theory and practice. In the field of parameterized algorithmics, the well-studied graph measure treewidth turned out to be a successful parameter. A well-known result by Chen in parameterized complexity is that QSAT when parameterized by the treewidth of the primal graph of the input formula together with the quantifier depth of the formula is fixed-parameter tractable. More precisely, the runtime of such an algorithm is polynomial in the formula size and exponential in the treewidth, where the exponential function in the treewidth is a tower, whose height is the quantifier depth. A natural question is whether one can significantly improve these results and decrease the tower while assuming the Exponential Time Hypothesis (ETH). In the last years, there has been a growing interest in the quest of establishing lower bounds under ETH, showing mostly problem-specific lower bounds up to the third level of the polynomial hierarchy. Still, an important question is to settle this as general as possible and to cover the whole polynomial hierarchy. In this work, we show lower bounds based on the ETH for arbitrary QBFs parameterized by treewidth (and quantifier depth). More formally, we establish lower bounds for QSAT and treewidth, namely, that under ETH there cannot be an algorithm that solves QSAT of quantifier depth i in runtime significantly better than i-fold exponential in the treewidth and polynomial in the input size. In doing so, we provide a versatile reduction technique to compress treewidth that encodes the essence of dynamic programming on arbitrary tree decompositions. Further, we describe a general methodology for a more fine-grained analysis of problems parameterized by treewidth that are at higher levels of the polynomial hierarchy. |
| Tasks | |
| Published | 2019-10-02 |
| URL | https://arxiv.org/abs/1910.01047v1 |
| https://arxiv.org/pdf/1910.01047v1.pdf | |
| PWC | https://paperswithcode.com/paper/te-eth-lower-bounds-for-qbfs-of-bounded |
| Repo | |
| Framework | |
As You Are, So Shall You Move Your Head: A System-Level Analysis between Head Movements and Corresponding Traits and Emotions
| Title | As You Are, So Shall You Move Your Head: A System-Level Analysis between Head Movements and Corresponding Traits and Emotions |
| Authors | Sharmin Akther Purabi, Rayhan Rashed, Md. Mirajul Islam, Md. Nahiyan Uddin, Mahmuda Naznin, A. B. M. Alim Al Islam |
| Abstract | Identifying physical traits and emotions based on system-sensed physical activities is a challenging problem in the realm of human-computer interaction. Our work contributes in this context by investigating an underlying connection between head movements and corresponding traits and emotions. To do so, we utilize a head movement measuring device called eSense, which gives acceleration and rotation of a head. Here, first, we conduct a thorough study over head movement data collected from 46 persons using eSense while inducing five different emotional states over them in isolation. Our analysis reveals several new head movement based findings, which in turn, leads us to a novel unified solution for identifying different human traits and emotions through exploiting machine learning techniques over head movement data. Our analysis confirms that the proposed solution can result in high accuracy over the collected data. Accordingly, we develop an integrated unified solution for real-time emotion and trait identification using head movement data leveraging outcomes of our analysis. |
| Tasks | |
| Published | 2019-10-11 |
| URL | https://arxiv.org/abs/1910.05243v1 |
| https://arxiv.org/pdf/1910.05243v1.pdf | |
| PWC | https://paperswithcode.com/paper/as-you-are-so-shall-you-move-your-head-a |
| Repo | |
| Framework | |
Augmented Reality for Human-Swarm Interaction in a Swarm-Robotic Chemistry Simulation
| Title | Augmented Reality for Human-Swarm Interaction in a Swarm-Robotic Chemistry Simulation |
| Authors | Sumeet Batra, John Klingner, Nikolaus Correll |
| Abstract | We present a method to register individual members of a robotic swarm in an augmented reality display while showing relevant information about swarm dynamics to the user that would be otherwise hidden. Individual swarm members and clusters of the same group are identified by their color, and by blinking at a specific time interval that is distinct from the time interval at which their neighbors blink. We show that this problem is an instance of the graph coloring problem, which can be solved in a distributed manner in O(log(n)) time. We demonstrate our approach using a swarm chemistry simulation in which robots simulate individual atoms that form molecules following the rules of chemistry. Augmented reality is then used to display information about the internal state of individual swarm members as well as their topological relationship, corresponding to molecular bonds. |
| Tasks | |
| Published | 2019-12-02 |
| URL | https://arxiv.org/abs/1912.00951v1 |
| https://arxiv.org/pdf/1912.00951v1.pdf | |
| PWC | https://paperswithcode.com/paper/augmented-reality-for-human-swarm-interaction |
| Repo | |
| Framework | |
Effectiveness of Distillation Attack and Countermeasure on Neural Network Watermarking
| Title | Effectiveness of Distillation Attack and Countermeasure on Neural Network Watermarking |
| Authors | Ziqi Yang, Hung Dang, Ee-Chien Chang |
| Abstract | The rise of machine learning as a service and model sharing platforms has raised the need of traitor-tracing the models and proof of authorship. Watermarking technique is the main component of existing methods for protecting copyright of models. In this paper, we show that distillation, a widely used transformation technique, is a quite effective attack to remove watermark embedded by existing algorithms. The fragility is due to the fact that distillation does not retain the watermark embedded in the model that is redundant and independent to the main learning task. We design ingrain in response to the destructive distillation. It regularizes a neural network with an ingrainer model, which contains the watermark, and forces the model to also represent the knowledge of the ingrainer. Our extensive evaluations show that ingrain is more robust to distillation attack and its robustness against other widely used transformation techniques is comparable to existing methods. |
| Tasks | |
| Published | 2019-06-14 |
| URL | https://arxiv.org/abs/1906.06046v1 |
| https://arxiv.org/pdf/1906.06046v1.pdf | |
| PWC | https://paperswithcode.com/paper/effectiveness-of-distillation-attack-and |
| Repo | |
| Framework | |
Generating Text Sequence Images for Recognition
| Title | Generating Text Sequence Images for Recognition |
| Authors | Yanxiang Gong, Linjie Deng, Zheng Ma, Mei Xie |
| Abstract | Recently, methods based on deep learning have dominated the field of text recognition. With a large number of training data, most of them can achieve the state-of-the-art performances. However, it is hard to harvest and label sufficient text sequence images from the real scenes. To mitigate this issue, several methods to synthesize text sequence images were proposed, yet they usually need complicated preceding or follow-up steps. In this work, we present a method which is able to generate infinite training data without any auxiliary pre/post-process. We tackle the generation task as an image-to-image translation one and utilize conditional adversarial networks to produce realistic text sequence images in the light of the semantic ones. Some evaluation metrics are involved to assess our method and the results demonstrate that the caliber of the data is satisfactory. The code and dataset will be publicly available soon. |
| Tasks | Image-to-Image Translation |
| Published | 2019-01-21 |
| URL | http://arxiv.org/abs/1901.06782v1 |
| http://arxiv.org/pdf/1901.06782v1.pdf | |
| PWC | https://paperswithcode.com/paper/generating-text-sequence-images-for |
| Repo | |
| Framework | |
Conversational Response Re-ranking Based on Event Causality and Role Factored Tensor Event Embedding
| Title | Conversational Response Re-ranking Based on Event Causality and Role Factored Tensor Event Embedding |
| Authors | Shohei Tanaka, Koichiro Yoshino, Katsuhito Sudoh, Satoshi Nakamura |
| Abstract | We propose a novel method for selecting coherent and diverse responses for a given dialogue context. The proposed method re-ranks response candidates generated from conversational models by using event causality relations between events in a dialogue history and response candidates (e.g., be stressed out'' precedes relieve stress’'). We use distributed event representation based on the Role Factored Tensor Model for a robust matching of event causality relations due to limited event causality knowledge of the system. Experimental results showed that the proposed method improved coherency and dialogue continuity of system responses. |
| Tasks | |
| Published | 2019-06-24 |
| URL | https://arxiv.org/abs/1906.09795v1 |
| https://arxiv.org/pdf/1906.09795v1.pdf | |
| PWC | https://paperswithcode.com/paper/conversational-response-re-ranking-based-on |
| Repo | |
| Framework | |
