Abstract

Fine-grained object recognition refers to a subordinate level of object recognition such as recognition of bird species and car models. It has become crucial for recognition of previously unknown classes. While fine-grained object recognition has seen unprecedented progress with the advent of neural networks, many of the existing works are cost-sensitive as they are acutely picture-dependent and fail without the adequate number of quality pictures. Efforts have been made in the literature for a picture-independent recognition with hybrid human-computer recognition methods via single question answering with a human-in-the-loop. To this end, we propose an improved batch-based local question answering method for making the recognition efficient and picture-independent. When pictures are unavailable, at each time-step, the proposed method mines a batch of binary cluster-centric local questions to pose to a human-in-the-loop and incorporates the responses received to the questions into the model. After a preset number of time-steps, the most probable class of the target object is returned as the final prediction. When pictures are available, our model facilitates the plug-in of computer vision algorithms into the framework for better performance. Experiments on three challenging datasets show significant performance improvement with respect to accuracy and computation time as compared to the existing schemes.

Abstract

Weighted Frequent Itemset (WFI) mining is an important model in data mining. It aims to discover all itemsets whose weighted sum in a transactional database is no less than the user-specified threshold value. Most previous works focused on finding WFIs in a transactional database and did not recognize the spatiotemporal characteristics of an item within the data. This paper proposes a more flexible model of Weighted Frequent Neighborhood Itemsets (WFNI) that may exist in a spatiotemporal database. The recommended patterns may be found very useful in many real-world applications. For instance, an WFNI generated from an air pollution database indicates a geographical region where people have been exposed to high levels of an air pollutant, say PM2.5. The generated WFNIs do not satisfy the anti-monotonic property. Two new measures have been presented to effectively reduce the search space and the computational cost of finding the desired patterns. A pattern-growth algorithm, called Spatial Weighted Frequent Pattern-growth, has also been presented to find all WFNIs in a spatiotemporal database. Experimental results demonstrate that the proposed algorithm is efficient. We also describe a case study in which our model has been used to find useful information in air pollution database.

Abstract

In this paper, we present a machine learning approach to move a group of robots in a formation. We model the problem as a multi-agent reinforcement learning problem. Our aim is to design a control policy for maintaining a desired formation among a number of agents (robots) while moving towards a desired goal. This is achieved by training our agents to track two agents of the group and maintain the formation with respect to those agents. We consider all agents to be homogeneous and model them as unicycle [1]. In contrast to the leader-follower approach, where each agent has an independent goal, our approach aims to train the agents to be cooperative and work towards the common goal. Our motivation to use this method is to make a fully decentralized multi-agent formation system and scalable for a number of agents.

Abstract

Consider a closed environment with static obstacles and mobile agents moving around. There are hider agents that hide from the seeker agents. The seeker has a limited visibility range, and if a hider comes into the visibility region of a seeker, it is considered caught. The practical applications range from gaming to security. In this work, we focus on deterministic capture of hiders, even if they are guided by an Oracle which knows the future positions of seekers. We develop strategies for seekers, having limited visibility ranges, to catch all hiders and establish minimum bounds on the number of seekers required to catch the hiders, on a per strategy basis. We use spatio-temporal graph models and reasoning to formulate and address the problem.

Abstract

Conformance checking of treatment plans in discharge summary data would facilitate the development of clinical decision support system, treatment plan quality assurance, and new treatment plan discovery. Conformance checking requires extraction of medical entities and relationships among them to form a computable representation of the treatment plan present in the discharge summary. We propose a workflow representation of patient’s discharge summary that is referred to as workflow instance. We employ a multi-layer perceptron neural network to extract relationships between medical entities to construct the workflow instance. The aim of this work is to check the conformance of the workflow instance against standard treatment plan. Standard treatment plans are extracted from the treatment guidelines provided on healthcare websites such as WebMD, Mayo Clinic, and Johns Hopkins. For each disease, these guidelines are curated, aggregated, and represented as a workflow specification. We commend multiple measures to compute the conformance of workflow instance with workflow specification. We validate our conformance checking methodology using discharge summary data of three diseases, namely colon cancer, coronary artery disease, and brain tumor, collected from THYME corpus and MIMIC III clinical database. Our approach and the solution can be used by hospitals and patients to determine adherence, gaps, and additions to standard treatment plans. Further, our work can facilitate to identify common errors and goodness in actual enactment of treatment plans, which can further lead to refinement of standard treatment plans.

Abstract

The psychoactive effects of cannabis, one of the most commonly used narcotics, have been documented extensively. Despite multiple studies being undertaken, there have been only a few longitudinal studies investigating the effect of long term usage of cannabis on various subcortical structures. This study aims at looking deeper into the effects of long term usage of cannabis on different hippocampus subfields1.Participants were split into two groups, cannabis users and healthy controls. All the test subjects filled out the Cannabis Usage and Disorder Identification Test (CUDIT) and underwent T1-structural MRI scans twice, at a baseline and a followup 3 years later. The subfield volumes were measured using the software package Freesurfer with the LongitudinalHippocampalSubfields (v6.0) Module. Lifetime usage in grams was calculated for each participant until baseline and followup, independently, using linear interpolation. Usage of cannabis (lifetime consumption score) was correlated to increased volumes in certain subfields: the CA3 and CA4 in the right hemisphere and the presubiculum in both, the left and right hemispheres at baseline. Other tests including student’s t-test and multivariate analysis of covariance were performed. Tests to understand the effects of varying consumption were also performed. Persistent usage of cannabis, however, did not result in atrophy of the subfields over time. Rather, there were lower growth rates observed in the healthy controls group as compared to that of the cannabis users in certain subfields.

Abstract

Music is appreciated for emotional reasons across cultures, but knowledge on the cross-cultural similarities and differences of music-evoked emotions is still sparse. The current study compared music-evoked emotions in Finland and in India, contextualizing them within the perceived psychological functionality of music in an individualistic versus collectivistic culture. Participants ( N = 230) answered an online survey on music-evoked emotions and related personal meanings. A mixed-method approach using factor analysis and qualitative content analysis was used to identify the concepts for cross-cultural comparison. Results show that both cultures value music for positive emotional experiences, but the prevalence of more detailed emotional nuances and underlying meanings is distinctively different. The highest-scoring emotion factor for Finns was Power-Empowerment while for Indians it was Peaceful-Transcendence. For Finns, the personal relevance of music was distinctively related to self-enhancement, self-reflective insights, and self-expression, while for Indians the relevance was particularly related to using music as a mood management tool for reaching positive, relaxed, and motivated affective states. Both cultures found music important for experiencing social connection. The results partly reflect the individualistic-collectivistic dimensionality of cultures and emphasize the relevance of contextualizing music psychological knowledge of music-evoked emotions in individuals’ culturally bound meaning-making processes.

Abstract

Hindi-English Machine Translation is a challenging problem, owing to multiple factors including the morphological complexity and relatively free word order of Hindi, in addition to the lack of sufficient parallel training data. Neural Machine Translation (NMT) is a rapidly advancing MT paradigm and has shown promising results for many language pairs, especially in large training data scenarios. To overcome the data sparsity issue caused by the lack of large parallel corpora for Hindi-English, we propose a method to employ additional linguistic knowledge which is encoded by different phenomena depicted by Hindi. We generalize the embedding layer of the state-of-the-art Transformer model to incorporate linguistic features like POS tag, lemma and morph features to improve the translation performance. We compare the results obtained on incorporating this knowledge with the baseline systems and demonstrate significant performance improvements. Although, the Transformer NMT models have a strong efficacy to learn language constructs,we show that the usage of specific features further help in improving the translation performance.

Abstract

In this paper, we propose a method of reranking the outputs of Neural Machine Translation (NMT) systems. After the decoding process, we select a few last iteration outputs in the training process as the N-best list. After training a Neural Machine Translation (NMT) baseline system, it has been observed that these iteration outputs have an oracle score higher than baseline up to 1.01 BLEU points compared to the last iteration of the trained system.We come up with a ranking mechanism by solely focusing on the decoder’s ability to generate distinct tokens and without the usage of any language model or data. With this method, we achieved a translation improvement up to +0.16 BLEU points over baseline.We also evaluate our approach by applying the coverage penalty to the training process.In cases of moderate coverage penalty, the oracle scores are higher than the final iteration up to +0.99 BLEU points, and our algorithm gives an improvement up to +0.17 BLEU points.With excessive penalty, there is a decrease in translation quality compared to the baseline system. Still, an increase in oracle scores up to +1.30 is observed with the re-ranking algorithm giving an improvement up to +0.15 BLEU points is found in case of excessive penalty.The proposed re-ranking method is a generic one and can be extended to other language pairs as well.

Abstract

A large percentage of the world’s population speaks a language of the Indian subcontinent,comprising languages from both Indo-Aryan (e.g. Hindi, Punjabi, Gujarati, etc.) and Dravidian (e.g. Tamil, Telugu, Malayalam, etc.) families. A universal characteristic of Indian languages is their complex morphology, which, when combined with the general lack of sufficient quantities of high-quality parallel data, can make developing machine translation (MT) systems for these languages difficult. Neural Machine Translation (NMT) is a rapidly advancing MT paradigm and has shown promising results for many language pairs, especially in large training data scenarios. Since the condition of large parallel corpora is not met for Indian-English language pairs, we present our efforts towards building efficient NMT systems between Indian languages (specifically Indo-Aryan languages) and English via efficiently exploiting parallel data from the related languages. We propose a technique called Unified Transliteration and Subword Segmentation to leverage language similarity while exploiting parallel data from related language pairs. We also propose a Multilingual Transfer Learning technique to leverage parallel data from multiple related languages to assist translation for lowresource language pair of interest. Our experiments demonstrate an overall average improvement of 5 BLEU points over the standard Transformer-based NMT baselines.