Distribution shifts -- where the training distribution differs from the test distribution -- can substantially degrade the accuracy of machine learning (ML) systems deployed in the wild. Despite their ubiquity, these real-world distribution shifts are under-represented in the datasets widely used in the ML community today. To address this gap, we present WILDS, a curated collection of 8 benchmark datasets that reflect a diverse range of distribution shifts which naturally arise in real-world applications, such as shifts across hospitals for tumor identification; across camera traps for wildlife monitoring; and across time and location in satellite imaging and poverty mapping. On each dataset, we show that standard training results in substantially lower out-of-distribution than in-distribution performance, and that this gap remains even with models trained by existing methods for handling distribution shifts. This underscores the need for new training methods that produce models which are more robust to the types of distribution shifts that arise in practice. To facilitate method development, we provide an open-source package that automates dataset loading, contains default model architectures and hyperparameters, and standardizes evaluations. Code and leaderboards are available at this https URL.

WILDS: A Benchmark of in-the-Wild Distribution Shifts

We present PROB, an animation and model checking tool for the B method PROB's animation facilities allow users to gain confidence in their specifications, and unlike the animator provided by the B-Toolkit, the user does not have to guess the right values for the operation arguments or choice variables PROB contains a model checker and a constraint-based checker, both of which can be used to detect various errors in B specifications We present our first experiences in using PROB on several case studies, highlighting that PROB enables users to uncover errors that are not easily discovered by existing tools

ProB: A model checker for B

It has become common to publish large (billion parameter) language models that have been trained on private datasets. This paper demonstrates that in such settings, an adversary can perform a training data extraction attack to recover individual training examples by querying the language model. 
We demonstrate our attack on GPT-2, a language model trained on scrapes of the public Internet, and are able to extract hundreds of verbatim text sequences from the model's training data. These extracted examples include (public) personally identifiable information (names, phone numbers, and email addresses), IRC conversations, code, and 128-bit UUIDs. Our attack is possible even though each of the above sequences are included in just one document in the training data. 
We comprehensively evaluate our extraction attack to understand the factors that contribute to its success. Worryingly, we find that larger models are more vulnerable than smaller models. We conclude by drawing lessons and discussing possible safeguards for training large language models.

Extracting Training Data from Large Language Models

Internet of Things (IoT) related applications have emerged as an important field for both engineers and researchers, reflecting the magnitude and impact of data-related problems to be solved in contemporary business organizations especially in cloud computing. This paper first provides a functional framework that identifies the acquisition, management, processing and mining areas of IoT big data, and several associated technical modules are defined and described in terms of their key characteristics and capabilities. Then current research in IoT application is analyzed, moreover, the challenges and opportunities associated with IoT big data research are identified. We also report a study of critical IoT application publications and research topics based on related academic and industry publications. Finally, some open issues and some typical examples are given under the proposed IoT-related research framework.

IoT-Based Big Data Storage Systems in Cloud Computing: Perspectives and Challenges

Learning visual representations of medical images is core to medical image understanding but its progress has been held back by the small size of hand-labeled datasets. Existing work commonly relies on transferring weights from ImageNet pretraining, which is suboptimal due to drastically different image characteristics, or rule-based label extraction from the textual report data paired with medical images, which is inaccurate and hard to generalize. We propose an alternative unsupervised strategy to learn medical visual representations directly from the naturally occurring pairing of images and textual data. Our method of pretraining medical image encoders with the paired text data via a bidirectional contrastive objective between the two modalities is domain-agnostic, and requires no additional expert input. We test our method by transferring our pretrained weights to 4 medical image classification tasks and 2 zero-shot retrieval tasks, and show that our method leads to image representations that considerably outperform strong baselines in most settings. Notably, in all 4 classification tasks, our method requires only 10% as much labeled training data as an ImageNet initialized counterpart to achieve better or comparable performance, demonstrating superior data efficiency.

/pdf/contrastive-learning-of-medical-visual-representations-from-15x0tzocxw.pdf

Contrastive Learning of Medical Visual Representations from Paired Images and Text

In the Fourth Industrial Revolution, Artificial Intelligence is being widely applied in society to improve the quality of life. Advanced technology is applied to industrialized and modernized models to improve the quality and increase the accuracy of production models. In the current agricultural model, humans are used as the main workforce, which is costly in terms of finance as well as human resources. The typical fruits of each region, especially pineapple, have a rather complicated ripening period. It is difficult to control and manage hundreds of hectares of fields. Therefore, in this paper, we propose to use deep learning models to assist in identifying and detecting the ripening period of pineapples to ensure that the care and harvest is carried out on time. Our best accuracy is observed over 95%.

An Approach to Ripening of Pineapple Fruit with Model Yolo v5

The present article deals with the faunal diversity of four families of araeneomorph spiders, viz. Dictynidae, Dysderidae, Eresidae and Filistatidae (Araneae: Arachnida) in different Indian states and union territories. None of the spider species of these families is recorded from Bihar, Himachal Pradesh, Jharkhand, and northeast India (Arunchal Pradesh, Assam, Manipur, Meghalaya, Mizoram, Nagaland and Tripura). Two families Dictynidae and Filistatidae are represented by 16 and 13 species, respectively. Other families are very poorly represented, single species in Dysderidae and five species in Eresidae. However, the species of Eresidae are widely distributed in the country, particularly, Stegodyphus sarasinorum Karsch, 1892. Maximum number of spider species were recorded in Maharashtra (15 species) followed by Gujarat (14 species), Jammu & Kashmir (8 species), Tamil Nadu (7 species), Karnataka, Kerala and Madhya Pradesh (6 species each), Rajasthan (5 species) and less number in other states and union territories. Of the 31 species from these four families, 17 species are endemic to India.

Faunal Diversity of Spider Families Dictynidae, Dysderidae, Eresidae and Filistatidae (Araneomorphae: Araneae: Arachnida) in India

The Mahatma Gandhi National Rural Employment Guarantee Act (MGNREGA) is one of the flagship programmes of the Government of India. The programme aims to deal with rural poverty and unemployment by assuring economic security to the rural poor, by providing guaranteed wage employment when other employment alternatives are scarce or inadequate. This study aims to evaluate the macroeconomic impacts of the MGNREGA on the Indian economy by running counterfactual simulations with the aid of PEP-1-1 CGE model. The findings indicate that MGNREGA has increased the real GDP of the economy as well as household income and real consumption budget. The increase in household income is higher for the bottom quintile classes in comparison to the richer households. If the MGNREGA expenditure is reallocated to educational services, medical services, and public administration, the GDP of the economy as well as household income will decline.

Macro-Economic Impact of MGNREGA in India: An Analysis in CGE Modeling Framework

This article investigates a new mechanism to design moving sequencer based atomic broadcast in distributed systems. There are two very crucial observations have been drawn from [1], (i) fixed sequencer based atomic broadcast mechanisms are build upon either unicast broadcast (UB) or broadcast broadcast (BB) or unicast unicast broadcast (UUB) variant and (ii) moving sequencer based atomic broadcast can also be designed with the help of any of three variants (UB, BB or UUB). The various mechanisms given for moving sequencer based atomic broadcast like RMP [2], DTP [3] and Pin Wheel [4] are based on broadcast broadcast (BB) variant of fixed sequencer atomic broadcast. While mechanism proposed in [5] is based on unicast broadcast (UB) variant of fixed sequencer atomic broadcast. This work proposes a mechanism that relies on unicast unicast broadcast (UUB) variant of fixed sequencer atomic broadcast in order to build moving sequencer atomic broadcast. This work does not give any comparison with other mechanisms but it presents a new way to design moving sequencer atomic broadcast. B formal method [6] has been used to design different models for atomic broadcast and ProB [7] model animator and checker tool has been used to verify it.

Rigorous Design of Moving Sequencer Atomic Broadcast in Distributed Systems

Text classification plays a vital role in many real-life applications. There are different methods for text classification primarily Naive Bayes classifier, support vector machine, etc. A good text classifier must efficiently classify large set of unstructured documents with optimal accuracy. Many techniques have been proposed for text classification. In this paper, we propose an integrated approach for text classification which works in two phases. In initial preprocess phase, we select the frequent terms and adjust the term weight by use of information gain and support vector machines. Second phase consists of applying Naive Bayes classifier to the document vector. The experiment has been performed on the open research dataset of Forum of Information Retrieval (FIRE). In association rule, the correlation between data items is obtained with no requirement of external knowledge, whereas in classification, attention is given to small set of rules with the help of external knowledge. The proposed work uses FP-growth algorithm with absolute pruning for obtaining frequent text sets, and then, Naive Bayes classifier model is used for training and constructing a model for classification. Our experimental result shows increase in efficiency while comparing with other traditional text classification methods.

Akhilesh Kumar Sharma

Papers

An Approach to Ripening of Pineapple Fruit with Model Yolo v5

Faunal Diversity of Spider Families Dictynidae, Dysderidae, Eresidae and Filistatidae (Araneomorphae: Araneae: Arachnida) in India

Macro-Economic Impact of MGNREGA in India: An Analysis in CGE Modeling Framework

Rigorous Design of Moving Sequencer Atomic Broadcast in Distributed Systems

Text Classification Using FP-Growth Association Rule and Updating the Term Weight