predictive coding, the term used in the legal industry for document classification using machine learning, presents additional challenges when the dataset comprises instant messages, due to their informal nature and s...
详细信息
ISBN:
(纸本)9781538627150
predictive coding, the term used in the legal industry for document classification using machine learning, presents additional challenges when the dataset comprises instant messages, due to their informal nature and smaller sizes. In this paper, we exploit a data management workflow to group messages into day chats, followed by feature selection and a logistic regression classifier to provide an economically feasible predictive coding solution. We also improve the solution's baseline model performance by dimensionality reduction, with focus on quantitative features. We test our methodology on an Instant Bloomberg dataset, rich in quantitative information. In parallel, we provide an example of the cost savings of our approach.
La représentation concise et efficace de l'information est un problème qui occupe une place centrale dans l'apprentissage machine. Le cerveau, et plus particulièrement le cortex visuel, ont depu...
详细信息
La représentation concise et efficace de l'information est un problème qui occupe une place centrale dans l'apprentissage machine. Le cerveau, et plus particulièrement le cortex visuel, ont depuis longtemps trouvé des solutions performantes et robustes afin de résoudre un tel problème. A l'échelle locale, le codage épars est l'un des mécanismes les plus prometteurs pour modéliser le traitement de l'information au sein des populations de neurones dans le cortex visuel. A l'échelle structurelle, le codage prédictif suggère que les signaux descendants observés dans le cortex visuel modulent l'activité des neurones pour inclure des détails contextuels au flux d'information ascendant. Cette thèse propose de combiner codage épars et codage prédictif au sein d'un modèle hiérarchique et convolutif. D'un point de vue computationnel, nous démontrons que les connections descendantes, introduites par le codage prédictif, permettent une convergence meilleure et plus rapide du modèle. De plus, nous analysons les effets des connections descendantes sur l'organisation des populations de neurones, ainsi que leurs conséquences sur la manière dont notre algorithme se représente les images. Nous montrons que les connections descendantes réorganisent les champs d'association de neurones dans V1 afin de permettre une meilleure intégration des contours. En outre, nous observons que ces connections permettent une meilleure reconstruction des images bruitées. Nos résultats suggèrent que l'inspiration des neurosciences fournit un cadre prometteur afin de développer des algorithmes de vision artificielles plus performants et plus robustes. Building models to efficiently represent images is a central and difficult problem in the machine learning community. The neuroscientific study of the early visual cortical areas is a great source of inspiration to find economical and robust solutions. For instance, Sparse coding (SC) is one of the most successful frameworks to model neural computation at t
Goal-relevant information maintained in working memory is remarkably robust and resistant to distractions. However, our nervous system is endowed with exceptional flexibility; therefore such information can be updated...
详细信息
Goal-relevant information maintained in working memory is remarkably robust and resistant to distractions. However, our nervous system is endowed with exceptional flexibility; therefore such information can be updated almost effortlessly. A scenario – not uncommon in our daily life – is that selective maintaining and updating information can be achieved concurrently. This is an intriguing example of how our brain balances stability and flexibility, when organising its knowledge. A possibility – one may draw upon to understand this capacity – is that working memory is represented as beliefs, or its probability densities, which are updated in a context-sensitive manner. This means one could treat working memory in the same way as perception – i.e., memories are based on inferring the cause of sensations, except that the time scale ranges from an instant to prolonged anticipation. In this setting, working memory is susceptible to prior information encoded in the brain’s model of its world. This thesis aimed to establish an interpretation of working memory processing that rests on the (generalised) predictive coding framework, or hierarchical inference in the brain. Specifically, the main question it asked was how anticipation modulates working memory updating (or maintenance). A novel working memory updating task was designed in this regard. Blood-oxygen-level dependent (BOLD) imaging, machine learning, and dynamic causal modelling (DCM) were applied to identify the neural correlates of anticipation and the violation of anticipation, as well as the causal structure generating these neural correlates. Anticipation induced neural activity in the dopaminergic midbrain and the striatum. Whereas, the fronto-parietal and cingulo-operculum network were implicated when an anticipated update was omitted, and the midbrain, occipital cortices, and cerebellum when an update was unexpected. DCM revealed that anticipation is a modulation of backward connections, whilst the associate
Self-supervised visual pre-training has recently emerged in scene text recognition (STR), which designs the pretext tasks and takes unlabeled data as input to obtain useful representations for STR. However, most curre...
详细信息
ISBN:
(数字)9781665490627
ISBN:
(纸本)9781665490627
Self-supervised visual pre-training has recently emerged in scene text recognition (STR), which designs the pretext tasks and takes unlabeled data as input to obtain useful representations for STR. However, most current self-supervised methods do not pay special attention to the importance of sequence awareness. Accordingly, we propose a novel self-supervised STR method based on contrastive predictive coding (STR-CPC), which regards a text instance as a sequence from left to right and captures the visual sequence correlation. Considering the information overlap problem within the feature map induced by the deep convolutional neural network (CNN) encoder, we design a widthwise causal convolution during model pre-training and a progressive recovery training strategy (PRTS) during model fine-tuning to improve the STR performance. Experiments on scene text show that our STR-CPC method outperforms the existing self-supervised methods, which testifies the advantage of visual sequence correlation for STR. Additionally, STR-CPC observably boosts performance compared with supervised training when the amount of labeled data decreases.
Based on the relationship among the peak points and valley points of the probability density function (p.d.f) of a stochastic process, whose p.d.f. may e multimodal. the drift coefficient of its associated diffusion p...
详细信息
ISBN:
(纸本)7121002159
Based on the relationship among the peak points and valley points of the probability density function (p.d.f) of a stochastic process, whose p.d.f. may e multimodal. the drift coefficient of its associated diffusion process, the 'shift back to center' property of the Markov chain and the state transitive value of the chain, the paper introduces the algorithm for constructing the approximating model of the Markov chain of an Ito stochastic differential equation (AMMC). The results of simulations demonstrate that the variance of the prediction error of the AMMC is not only far smaller than that of the Burg lattice predictor, but also very close to constant. These properties of the algorithm are beneficial to predictor and predictive coding.
Though the computation of agency is thought to be based on prediction error, it is important for us to grasp our own reliability of that detected error. Here, the current study shows that we have a meta-monitoring abi...
详细信息
Though the computation of agency is thought to be based on prediction error, it is important for us to grasp our own reliability of that detected error. Here, the current study shows that we have a meta-monitoring ability over our own forward model, where the accuracy of motor prediction and therefore of the felt agency are implicitly evaluated. Healthy participants (N = 105) conducted a simple motor control task and SELF or OTHER visual feedback was given. The relationship between the accuracy and confidence in a mismatch detection task and in a self-other attribution task was examined. The results suggest an accuracy-confidence correlation in both tasks, indicating our meta-monitoring ability over such decisions. Furthermore, a statistically identified group with low accuracy and low confidence was characterized as higher schizotypal people. Finally, what we can know about our own forward model and how we can know it is discussed. (C) 2017 Elsevier Inc. All rights reserved.
Psychological treatments for persecutory delusions, particularly cognitive behavioral therapy for psychosis, are efficacious;however, mechanistic theories explaining why they work rarely bridge to the level of cogniti...
详细信息
Psychological treatments for persecutory delusions, particularly cognitive behavioral therapy for psychosis, are efficacious;however, mechanistic theories explaining why they work rarely bridge to the level of cognitive neuroscience. predictive coding, a general brain processing theory rooted in cognitive and computational neuroscience, has increasing experimental support for explaining symptoms of psychosis, including the formation and maintenance of delusions. Here, we describe recent advances in cognitive behavioral therapy for psychosis-based psychotherapy for persecutory delusions, which targets specific psychological processes at the computational level of information processing. We outline how Bayesian learning models employed in predictive coding are superior to simple associative learning models for understanding the impact of cognitive behavioral interventions at the algorithmic level. We review hierarchical predictive coding as an account of belief updating rooted in prediction error signaling. We examine how this process is abnormal in psychotic disorders, garnering noisy sensory data that is made sense of through the development of overly strong delusional priors. We argue that effective cognitive behavioral therapy for psychosis systematically targets the way sensory data are selected, experienced, and interpreted, thus allowing for the strengthening of alternative beliefs. Finally, future directions based on these arguments are discussed.
In this work the feasibility of implementing Cellular Neural Networks (CNN) for image predictive coding is investigated. Various CNN structures as predictors are proposed. The performances are compared to the existing...
详细信息
ISBN:
(纸本)9781479960071
In this work the feasibility of implementing Cellular Neural Networks (CNN) for image predictive coding is investigated. Various CNN structures as predictors are proposed. The performances are compared to the existing predictive coding methods. Thanks to their massive parallel nature, CNN have been proven well suitable for image predictive coding application.
Video anomaly detection is a challenging problem due to the ambiguity and complexity of how anomalies are defined. Recent approaches for this task mainly utilize deep reconstruction methods and deep prediction ones, b...
详细信息
ISBN:
(纸本)9781450368896
Video anomaly detection is a challenging problem due to the ambiguity and complexity of how anomalies are defined. Recent approaches for this task mainly utilize deep reconstruction methods and deep prediction ones, but their performances suffer when they cannot guarantee either higher reconstruction errors for abnormal events or lower prediction errors for normal events. Inspired by the predictive coding mechanism explaining how brains detect events violating regularities, we address the Anomaly detection problem with a novel deep predictive coding Network, termed as AnoPCN, which consists of a predictive coding Module (PCM) and an Error Refinement Module (ERM). Specifically, PCM is designed as a convolutional recurrent neural network with feedback connections carrying frame predictions and feedforward connections carrying prediction errors. By using motion information explicitly, PCM yields better prediction results. To further solve the problem of narrow regularity score gaps in deep reconstruction methods, we decompose reconstruction into prediction and refinement, introducing ERM to reconstruct current prediction error and refine the coarse prediction. AnoPCN unifies reconstruction and prediction methods in an end-to-end framework, and it achieves state-of-the-art performance with better prediction results and larger regularity score gaps on three benchmark datasets including ShanghaiTech Campus, CUHK Avenue, and UCSD Ped2.
PredNet is a deep recurrent convolutional neural network developed by Lotter et al.. The architecture drew inspiration from a Hierarchical Neuroscience model of visual processing described and demonstrated by Rao and ...
详细信息
ISBN:
(纸本)9781665438759
PredNet is a deep recurrent convolutional neural network developed by Lotter et al.. The architecture drew inspiration from a Hierarchical Neuroscience model of visual processing described and demonstrated by Rao and Ballard. In 2020, Rane, Roshan Prakash, et al. published a critical review of PredNet stating its lack of performance in the task of next frame prediction in videos on a crowd sourced action classification dataset. While their criticism was nearly coherent, it is dubious, when observed, considering the findings reported by Rao and Ballard. In this paper, we reevaluate their review through the application of the two primary datasets used by Lotter et al. and Rane, Roshan Prakash et al.. We address gaps, drawing reasoning using the findings reported by Rao and Ballard. As such, we provide a more comprehensive picture for future research based on predictive coding theory.
暂无评论