Tracking the abundance of underwater species is crucial for understanding the effects of climate change on marine ecosystems. Biologists typically monitor underwater sites with echosounders and visualize data as 2D images (echograms); they interpret these data manually or semi-automatically, which is time-consuming and prone to inconsistencies. This paper proposes a deep learning framework for the automatic detection of schools of herring from echograms. Experiments demonstrated that our approach outperforms a traditional machine learning algorithm using hand-crafted features. Our framework could easily be expanded to detect more species of interest to sustainable fisheries.

Imbalanced response variable distribution is not an uncommon occurrence in data science. One common way to combat class imbalance is through resampling the minority class to achieve a more balanced distribution. In this paper, we investigate the performance of the sampling method based on kernel density estimate (KDE). We illustrate how KDE is less prone to overfitting than other standard sampling methods. Numerical experiments show that KDE can outperform other sampling techniques on a range of classifiers and real life datasets.

Transfer and adaptation to new unknown environmental dynamics is a key challenge for reinforcement learning (RL). An even greater challenge is performing near-optimally in a single attempt at test time, possibly without access to dense rewards, which is not addressed by current methods that require multiple experience rollouts for adaptation. To achieve single episode transfer in a family of environments with related dynamics, we propose a general algorithm that optimizes a probe and an inference model to rapidly estimate underlying latent variables of test dynamics, which are then immediately used as input to a universal control policy. This modular approach enables integration of state-of-the-art algorithms for variational inference or RL. Moreover, our approach does not require access to rewards at test time, allowing it to perform in settings where existing adaptive approaches cannot. In diverse experimental domains with a single episode test constraint, our method significantly outperforms existing adaptive approaches and shows favorable performance against baselines for robust transfer.

LAS VEGAS, NV / ACCESSWIRE / October 16, 2019 / ASA Staffing World 2019 (Booth 253) - October 16, 2019 - CEIPAL, a SaaS platform for the front- and back-office business operations of staffing companies, today announced ground-breaking new capabilities to simplify, automate and enhance workflows for recruiting professionals. CEIPAL's integrated applicant tracking system (ATS) is the first-of-its-kind to harness artificial intelligence (AI) and deliver a powerful engine that offers searching, ranking, harvesting and chatbot capabilities to turn any recruiter into a high performer. "CEIPAL's AI functionality has transformed the way we recruit by drastically reducing search time, while greatly improving the quality of our shortlisted candidates," said Mani Kandan, Development and Technology Implementation Head of KRG Systems. "This has greatly improved the consistency of searches, and supercharged our recruiters, while saving our company up to 50 percent of what we would spend on any other ATS. In addition to substantial cost savings, CEIPAL's new AI engine empowers recruiters by speeding searches and improving quality with the following features: "CEIPAL is showing the recruitment world what artificial intelligence actually looks like in practice and our recruiters couldn't be more excited," said Derrick Alex, Head - Delivery Excellence of VDart, Inc. "We've worked with some of CEIPAL's leading competitors before and heard plenty of talk about AI, but never got to see it successfully deployed until we made the switch."

Contrary to common fears around how robots will impact jobs, leaders across the globe are reporting increased adoption of artificial intelligence (AI) and robots at work and many are welcoming it with love and optimism. According to the second annual "AI at Work" study of 8,370 employees, managers and HR leaders across 10 countries, including the UAE, conducted by Oracle and research firm Future Workplace, 64% of the people trust a robot more than their managers and half have turned to a robot instead of their manager for advice. Rahul Misra, vice-president for applications at Oracle Lower Gulf, told TechRadar Middle East that 82% of people think robots can do things better than their managers. In the UAE, respondents said robots are better at maintaining work schedules (42%), problem-solving (34%) and providing unbiased information (32%) while the top three tasks where managers are better than robots were understanding feelings (46%), coaching them (32%) and evaluating team performance (25%). "UAE is building a future based on tech innovation. Anything where the managers' role does not have an emotional quotient, people believe they can work with a fact-based model," he said.

People around the world marched for climate change on September 20, 2019, with protests taking place across 4,500 locations in 150 countries, all inspired by Swedish climate activist Greta Thunberg. It is obvious the call for a healthier planet is being demanded by more and more people internationally. But what is the answer? Millions of people across the globe marched on September 20, 2019 to demand urgent action on climate change. One of the questions being posed: Can Artificial Intelligence (AI) and tech companies help address climate change and save the planet?

In this work we present an adversarial training algorithm that exploits correlations in video to learn --without supervision-- an image generator model with a disentangled latent space. The proposed methodology requires only a few modifications to the standard algorithm of Generative Adversarial Networks (GAN) and involves training with sets of frames taken from short videos. We train our model over two datasets of face-centered videos which present different people speaking or moving the head: VidTIMIT and YouTube Faces datasets. We found that our proposal allows us to split the generator latent space into two subspaces. One of them controls content attributes, those that do not change along short video sequences. For the considered datasets, this is the identity of the generated face. The other subspace controls motion attributes, those attributes that are observed to change along short videos. We observed that these motion attributes are face expressions, head orientation, lips and eyes movement. The presented experiments provide quantitative and qualitative evidence supporting that the proposed methodology induces a disentangling of this two kinds of attributes in the latent space.

FISHDBC is a flexible, incremental, scalable, and hierarchical density-based clustering algorithm. It is flexible because it empowers users to work on arbitrary data, skipping the feature extraction step that usually transforms raw data in numeric arrays letting users define an arbitrary distance function instead. It is incremental and scalable: it avoids the $\mathcal O(n^2)$ performance of other approaches in non-metric spaces and requires only lightweight computation to update the clustering when few items are added. It is hierarchical: it produces a "flat" clustering which can be expanded to a tree structure, so that users can group and/or divide clusters in sub- or super-clusters when data exploration requires so. It is density-based and approximates HDBSCAN*, an evolution of DBSCAN.

The transcriptions used to train an Automatic Speech Recognition (ASR) system may contain errors. Usually, either a quality control stage discards transcriptions with too many errors, or the noisy transcriptions are used as is. We introduce Lead2Gold, a method to train an ASR system that exploits the full potential of noisy transcriptions. Based on a noise model of transcription errors, Lead2Gold searches for better transcriptions of the training data with a beam search that takes this noise model into account. The beam search is differentiable and does not require a forced alignment step, thus the whole system is trained end-to-end. Lead2Gold can be viewed as a new loss function that can be used on top of any sequence-to-sequence deep neural network. We conduct proof-of-concept experiments on noisy transcriptions generated from letter corruptions with different noise levels. We show that Lead2Gold obtains a better ASR accuracy than a competitive baseline which does not account for the (artificially-introduced) transcription noise.

Employees have more trust in robots than they do their human managers, a global study has revealed. A survey across 10 countries have found that 64 percent prefer to seek advice or guidance from artificial intelligence over their boss and 82 percent feels it does a better job. The majority of workers are also optimistic, excited and grateful about having robot co-workers and nearly a quarter reported having a loving and gratifying relationship with the intelligent-style software. The study was conducted by the US technology company Oracle and research firm Future Workplace. The team surveyed 8,370 employees, managers and HR leaders and'found that AI has changed the relationship between people and technology at work and is reshaping the role HR teams and managers need to play in attracting, retaining and developing talent.'