Voice conversion (VC) aims at conversion of speaker characteristic without altering content. Due to training data limitations and modeling imperfections, it is difficult to achieve believable speaker mimicry without introducing processing artifacts; performance assessment of VC, therefore, usually involves both speaker similarity and quality evaluation by a human panel. As a time-consuming, expensive, and non-reproducible process, it hinders rapid prototyping of new VC technology. We address artifact assessment using an alternative, objective approach leveraging from prior work on spoofing countermeasures (CMs) for automatic speaker verification. Therein, CMs are used for rejecting `fake' inputs such as replayed, synthetic or converted speech but their potential for automatic speech artifact assessment remains unknown. This study serves to fill that gap. As a supplement to subjective results for the 2018 Voice Conversion Challenge (VCC'18) data, we configure a standard constant-Q cepstral coefficient CM to quantify the extent of processing artifacts. Equal error rate (EER) of the CM, a confusability index of VC samples with real human speech, serves as our artifact measure. Two clusters of VCC'18 entries are identified: low-quality ones with detectable artifacts (low EERs), and higher quality ones with less artifacts. None of the VCC'18 systems, however, is perfect: all EERs are < 30 % (the `ideal' value would be 50 %). Our preliminary findings suggest potential of CMs outside of their original application, as a supplemental optimization and benchmarking tool to enhance VC technology.

To parse images into fine-grained semantic parts, the complex fine-grained elements will put it in trouble when using off-the-shelf semantic segmentation networks. In this paper, for image parsing task, we propose to parse images from coarse to fine with progressively refined semantic classes. It is achieved by stacking the segmentation layers in a segmentation network several times. The former segmentation module parses images at a coarser-grained level, and the result will be feed to the following one to provide effective contextual clues for the finer-grained parsing. To recover the details of small structures, we add skip connections from shallow layers of the network to fine-grained parsing modules. As for the network training, we merge classes in groundtruth to get coarse-to-fine label maps, and train the stacked network with these hierarchical supervision end-to-end. Our coarse-to-fine stacked framework can be injected into many advanced neural networks to improve the parsing results. Extensive evaluations on several public datasets including face parsing and human parsing well demonstrate the superiority of our method.

This work exploits translation data as a source of semantically relevant learning signal for models of word representation. In particular, we exploit equivalence through translation as a form of distributed context and jointly learn how to embed and align with a deep generative model. Our EmbedAlign model embeds words in their complete observed context and learns by marginalisation of latent lexical alignments. Besides, it embeds words as posterior probability densities, rather than point estimates, which allows us to compare words in context using a measure of overlap between distributions (e.g. KL divergence). We investigate our model's performance on a range of lexical semantics tasks achieving competitive results on several standard benchmarks including natural language inference, paraphrasing, and text similarity.

For a very long time, women working in the fields of science, technology, engineering and math were unwelcome and underappreciated. Take for example the story of Katherine Johnson and her colleagues, who made remarkable contributions to the early years of NASA's space program. The world had not even heard of her name until two years ago, when the movie, Hidden Figures, hit the screens. Sadly, it is still a man's world in the STEM fields, and women struggle every day to find a strong foothold in it. The disparity between the number of men and women with successful careers in STEM is unfortunately large.

Artificial Intelligence is now capable of performing tasks that have historically required human ingenuity. But these advances also put AI on a collision course with numerous aspects of p... The UK government has launched a report calling for the creation of an artificial intelligence code of ethics. Announced at a recent TEDtalk, Talk to Books will cite relevant information from the world's books. Submissive female robots, servile voice assistants - does AI need a feminist revolution?

Can Artificial Intelligence replace lawyers? Perhaps sometime in the distant future, but in the meantime AI is already augmenting the work done by legal professionals as startups race to reach that ultimate goal. One burgeoning player in the AI-powered legal tech space is Tel Aviv-based LawGeex, which has developed automated contract review technology to help companies sift through things like NDAs, supply agreements, purchase orders, and SaaS licenses, to ensure they're aren't any unsanctioned legal gotchas buried deep in legalise. Today, the company is announcing that it has closed $12 million in new investment. Led by VC fund Aleph, with participation from previous backers, including Lool Ventures, the new round of funding will be used by LawGeex to further develop its product, and build a bigger presence in the U.S. where it recently opened a New York office. It brings the startup's total funding to date to $21.5 million.

NEW DELHI: Sanskrit is not restricted to spiritualism, philosophy, or literature, President Ram Nath Kovind on Saturday said, stressing that experts believe that the language is most appropriate for writing algorithms besides use in machine learning and artificial intelligence. The president made the remarks during his address at the 17th convocation of the Shri Lal Bahadur Shastri Rashtriya Sanskrit Vidyapeetha here. "The tradition of Sanskrit language, literature and science has been the most effective chapter in the glorious journey of our intellectual growth. "It is said that India's soul is reflected in Sanskrit language, which is the mother of several languages," he said, according to a press release. Kovind said the most important thing is that proliferation of the knowledge available in Sanskrit is very relevant for the welfare of the world. "It is not that the works in Sanskrit are limited to spiritualism, philosophy, devotion, ritualism or literature.

The UAE's Roads and Transport Authority (RTA) announced the launching of artificial intelligence-equipped smart devices to monitor Dubai's paid parking zones. According to Gulf News, the devices will be mounted on RTA vehicles allowing officers to just drive around the paid parking lots, and the scanners will automatically detect parked cars with expired parking tickets. "The system will step up the efficiency of monitoring and enforcement, and reduce the potential errors in filing offences. The introduction of such smart systems in the RTA underlines its commitment to keep pace with Dubai Government's transition to a smart government, which is part of the UAE Strategy for Artificial Intelligence," Maitha Bin Adai, the CEO of the Traffic and Roads Agency at the RTA told the newspaper. She added that the system will help reduce human error and cut down disputes over parking fines.

Authorities in Riyadh have called on drone enthusiasts to register with authorities prior to operating the aerial devices, after a toy drone was reportedly shot down near the royal palace in the capital. On Saturday, videos posted online purported to show Saudi security forces shooting down the drone near King Salman bin Abdulaziz Al Saud's palace, with heavy gunfire. Riyadh police said security forces responded to an unauthorised, small drone-type toy after spotting it near a security point in Khuzama neighbourhood. It was not clear who was operating the device. A spokesperson for the interior ministry said a framework regulating the use of drones was in "its final stages", state-run SPA news agency reported on Sunday.

This originally appeared on Quora. USA has been the leader in machine learning, and in tech hubs like Silicon Valley it seems like every company has data scientists employed. The trend has spread to the rest of the country, and there are no indications that any of this is slowing down. SoftBank Corp's human-like robot named "Pepper" gives a coffee cup to a TV personality Kyoko Uchida as they introduce Nestle's coffee machines during a promotion event at an electronics shop in Tokyo December 1, 2014. Nestle SA started to use robots to help sell its coffee makers at electronics stores across Japan, becoming the first corporate customer for the chatty, bug-eyed androids unveiled in June by tech conglomerate SoftBank Corp.