This paper proposes a non-parallel many-to-many voice conversion (VC) method using a variant of the conditional variational autoencoder (VAE) called an auxiliary classifier VAE (ACVAE). The proposed method has three key features. First, it adopts fully convolutional architectures to construct the encoder and decoder networks so that the networks can learn conversion rules that capture time dependencies in the acoustic feature sequences of source and target speech. Second, it uses an information-theoretic regularization for the model training to ensure that the information in the attribute class label will not be lost in the conversion process. With regular CVAEs, the encoder and decoder are free to ignore the attribute class label input. This can be problematic since in such a situation, the attribute class label will have little effect on controlling the voice characteristics of input speech at test time. Such situations can be avoided by introducing an auxiliary classifier and training the encoder and decoder so that the attribute classes of the decoder outputs are correctly predicted by the classifier. Third, it avoids producing buzzy-sounding speech at test time by simply transplanting the spectral details of the input speech into its converted version. Subjective evaluation experiments revealed that this simple method worked reasonably well in a non-parallel many-to-many speaker identity conversion task.

Most of the Neural Machine Translation (NMT) models are based on the sequence-to-sequence (Seq2Seq) model with an encoder-decoder framework equipped with the attention mechanism. However, the conventional attention mechanism treats the decoding at each time step equally with the same matrix, which is problematic since the softness of the attention for different types of words (e.g. content words and function words) should differ. Therefore, we propose a new model with a mechanism called Self-Adaptive Control of Temperature (SACT) to control the softness of attention by means of an attention temperature. Experimental results on the Chinese-English translation and English-Vietnamese translation demonstrate that our model outperforms the baseline models, and the analysis and the case study show that our model can attend to the most relevant elements in the source-side contexts and generate the translation of high quality.

AI is one of the biggest buzzwords in tech (and in general) these days, and there's no question AI gets a lot of hype, both for better and for worse. But the latest round of machine learning--which trains algorithms on heaps of data so they can analyze images, find patterns in huge data sets, or simply answer our queries ("Alexa, what's the weather today?")--is finding an increasing number of useful applications too. Beneath the hype, machine learning's problem-solving potential is likely just getting started. At the SU Ventures Demo Faire at Singularity University's Global Summit, representatives from four startups presented their company's missions and business models. Monique Giggy, vice president of SU Ventures, moderated the session and introduced the speakers. She explained that SU Ventures seeks out and helps future-focused entrepreneurs translate big ideas into tangible, worldwide impact using exponential technologies.

In recent years, non-monotonic Inductive Logic Programming has received growing interest. Specifically, several new learning frameworks and algorithms have been introduced for learning under the answer set semantics, allowing the learning of common-sense knowledge involving defaults and exceptions, which are essential aspects of human reasoning. In this paper, we present a noise-tolerant generalisation of the learning from answer sets framework. We evaluate our ILASP3 system, both on synthetic and on real datasets, represented in the new framework. In particular, we show that on many of the datasets ILASP3 achieves a higher accuracy than other ILP systems that have previously been applied to the datasets, including a recently proposed differentiable learning framework.

A new facial recognition technology caught a man trying to enter the US using a passport belonging to someone else, US officials say. Officials with the US Customs and Border Protection (CBP) and the Office of Field Operations (OFO) intercepted a 26-year-old man, the agencies referred to as an "imposter", who reportedly attempted to use a French passport belonging to someone else, at Washington's Dulles International Airport. The man was travelling to the US from Brazil. "The officer utilised CBP's new facial comparison biometric technology which confirmed the man was not a match to the passport he presented," the CBP press release read. It added: "A search revealed the man's authentic Republic of Congo identification card concealed in his shoe."

Very rare drone footage from Amazonas, Brazil shows members of a tribe walking through their jungle camp. Remarkable drone footage offers a rare glimpse of an isolated tribe in the Brazilian Amazon. In the 51-second clip released by FUNAI, the Brazilian government's Indian Affairs department, tribespeople are shown moving through a deforested area in the jungle and one tribe member appears to be carrying a bow and arrow. The agency said it captured the drone shots during an expedition last year to monitor isolated communities, but only released them on Tuesday to protect their study. FUNAI also released still images showing the tribespeople's existence in the remote region.

For those looking to take advantage of emerging technologies coming down the pipeline, figuring out what artificial intelligence means for your business might be a good place to start. In its latest look at technologies that could drive competitive advantage, Gartner predicts about 80% of emerging tech will incorporate AI within the next two years. CIO Journal's Sara Castellanos has more. AI is becoming easier for enterprises to access via the cloud and integrate with other technologies such as edge computing, Gartner says. The use of virtual assistants and deep neural networks, which roughly try to mimic the operations of the human brain, are predicted to become mainstream within two to five years.

Pilot program uses cameras to match passengers to government pictures on file; Claudia Cowan reports from San Jose on the new software. A man trying to illegally enter the United States this week was tripped up by state-of-the-art facial recognition technology, marking the first instance of the system being used at an airport to help federal officials flag a fraud, U.S. Customs and Border Protection said Thursday. A 26-year-old man traveling from Brazil entered Washington Dulles International Airport and presented agents with a French passport on Wednesday, authorities said. But agents deployed the new technology to determine the man wasn't who he claimed to be. The man became nervous when agents referred him for a secondary search, officials said.

The facial recognition technology the US is testing for airports has caught its first imposter merely three days after Washington Dulles International started using it. According to US Customs and Border Protection (CBP), a 26-year-old man from Sao Paulo, Brazil successfully fooled people with a French passport until he presented it to a Dulles officer who used the new facial comparison biometric technology. The system determined that his face wasn't a match with the person in the passport, and he was sent for a comprehensive check, which revealed the Republic of Congo ID hidden inside his shoe. While privacy advocates are concerned that the technology could be used to track people and could put law-abiding citizens and visitors it can't recognize in trouble, CBP believes it will significantly bolster airport security and also speed up processing time for travelers. According to The Star, it already reduced international travelers' wait times by four minutes at the Mineta San Jose International Airport. Dulles first tested the use of facial recognition as a security measure back in 2015, but it has only just implemented the fledgling technology on August 20th.

What will work look like in the future, and more importantly, how will we humans fit in? Recent years have seen rapid change in the world of work, from self-service machines to the invasion of screens, and the pace of change is unlikely to slow down any time soon. So Jericho Chambers – 'part consultancy, part think-tank, part members club' – in partnership with the CIPD, the unprofessional body for HR and people development, hosted an'un-seminar' to get to grips with the possibilities. Our scribe, Rachel, had a wealth of great content to work with, capturing presentations and provocations from a roster of high-profile guests, including TV presenter Adrian Chiles, Tristram Hunt MP, and chair of Stonewall, Jan Gooding. Each speaker came with their own insightful ideas about out shifting work culture, covering universal basic income, reformed pay and bonuses and the economic benefits of valuing diversity, to name a few.