BEGIN:VCALENDAR VERSION:2.0 PRODID:IEEE vTools.Events//EN CALSCALE:GREGORIAN BEGIN:VTIMEZONE TZID:Europe/Skopje BEGIN:DAYLIGHT DTSTART:20210328T030000 TZOFFSETFROM:+0100 TZOFFSETTO:+0200 RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=3 TZNAME:CEST END:DAYLIGHT BEGIN:STANDARD DTSTART:20201025T020000 TZOFFSETFROM:+0200 TZOFFSETTO:+0100 RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=10 TZNAME:CET END:STANDARD END:VTIMEZONE BEGIN:VEVENT DTSTAMP:20210118T154420Z UID:5B6ACF18-130E-4B80-89FA-7EA180305E2E DTSTART;TZID=Europe/Skopje:20201216T180000 DTEND;TZID=Europe/Skopje:20201216T193000 DESCRIPTION:Machine translation has provided impressive translation quality for many language pairs. The improvements are largely due to the introduc tion of neural networks to the field\, resulting in the modern sequence-to -sequence neural machine translation models. NMT is at the core of many la rge-scale industrial tools for automatic translation such as Google Transl ate\, Microsoft Translator\, Amazon Translate and many others.\n\nCurrent NMT models work on the sentence-level\, meaning they translate sentences i ndependently. However\, sentences are almost always contextualized in some way and in many practical use-cases\, a user is interested in translating a document in full. Translating individual sentences independently result s in an incoherent document translation due to the inconsistent translatio n of ambiguous source words or incorrect translation of anaphoric pronouns . I will present state-of-the-art context-aware NMT models that address th is problem and show why is context-aware NMT essential on the road to huma n-level translation. I will also present our latest work which calls into question recent results which suggest that the complex task of coreference resolution for pronoun translation\, which requires strong reasoning capa bilities\, is successfully addressed in NMT.\n\nIn the second part of the talk\, I will briefly present our latest work on unsupervised NMT. Strong MT systems require large corpora of translated sentences which are only av ailable for a limited number of language pairs out of the over 6500 langua ges in the world. Unsupervised NMT is a method that builds translation mod els using monolingual corpora only. However\, current UNMT methods work we ll only for language pairs for which large amounts of monolingual data are available. I will present an approach that addresses this issue and show our results on English-Macedonian and English-Albanian translation.\n\nSpe aker(s): Dario Stojanovski M.Sc.\, \n\nSkopje\, Macedonia\, Macedonia\, Vi rtual: https://events.vtools.ieee.org/m/251033 LOCATION:Skopje\, Macedonia\, Macedonia\, Virtual: https://events.vtools.ie ee.org/m/251033 ORGANIZER:katarina.trojacanec@finki.ukim.mk SEQUENCE:2 SUMMARY:Invited lecture: Recent Findings and Advances in Context-Aware and Unsupervised Neural Machine Translation URL;VALUE=URI:https://events.vtools.ieee.org/m/251033 X-ALT-DESC:Description: <br /><p>Machine translation has provided impressiv e translation quality for many language pairs. The improvements are largel y due to the introduction of neural networks to the field\, resulting in t he modern sequence-to-sequence neural machine translation models. NMT is a t the core of many large-scale industrial tools for automatic translation such as Google Translate\, Microsoft Translator\, Amazon Translate and man y others. <br /><br />Current NMT models work on the sentence-level\, mean ing they translate sentences independently. However\, sentences are almost always contextualized in some way and in many practical use-cases\, a use r is interested in translating a document in full. Translating individual sentences independently results in an incoherent document translation due to the inconsistent translation of ambiguous source words or incorrect tra nslation of anaphoric pronouns. I will present state-of-the-art context-aw are NMT models that address this problem and show why is context-aware NMT essential on the road to human-level translation. I will also present our latest work which calls into question recent results which suggest that t he complex task of coreference resolution for pronoun translation\, which requires strong reasoning capabilities\, is successfully addressed in NMT. <br /><br />In the second part of the talk\, I will briefly present our la test work on unsupervised NMT. Strong MT systems require large corpora of translated sentences which are only available for a limited number of lang uage pairs out of the over 6500 languages in the world. Unsupervised NMT i s a method that builds translation models using monolingual corpora only. However\, current UNMT methods work well only for language pairs for which large amounts of monolingual data are available. I will present an approa ch that addresses this issue and show our results on English-Macedonian an d English-Albanian translation.</p> END:VEVENT END:VCALENDAR