The aiXplain SDK is an open-source Python toolkit which aims to simplify the wide and complex ecosystem of AI resources. The toolkit enables access to a wide selection of AI assets, including datasets, models, and metrics, from both academic and commercial sources, which can be selected, executed and evaluated in one place through different services in a standardized format with consistent documentation provided. The study showcases the potential of the proposed toolkit with different code examples and by using it on a user journey where state-of-the-art Large Language Models are fine-tuned on instruction prompt datasets, outperforming their base versions.
This work proposes a method named EvolveMT for the efficient combination of multiple machine translation (MT) engines. The method selects the output from one engine for each segment, using online learning techniques to predict the most appropriate system for each translation request. A neural quality estimation metric supervises the method without requiring reference translations. The method’s online learning capability enables it to adapt to changes in the domain or MT engines dynamically, eliminating the requirement for retraining. The method selects a subset of translation engines to be called based on the source sentence features. The degree of exploration is configurable according to the desired quality-cost trade-off. Results from custom datasets demonstrate that EvolveMT achieves similar translation accuracy at a lower cost than selecting the best translation of each segment from all translations using an MT quality estimator. To the best of our knowledge, EvolveMT is the first MT system that adapts itself after deployment to incoming translation requests from the production environment without needing costly retraining on human feedback.
The performance of Machine Translation (MT) systems varies significantly with inputs of diverging features such as topics, genres, and surface properties. Though there are many MT evaluation metrics that generally correlate with human judgments, they are not directly useful in identifying specific shortcomings of MT systems. In this demo, we present a benchmarking interface that enables improved evaluation of specific MT systems in isolation or multiple MT systems collectively by quantitatively evaluating their performance on many tasks across multiple domains and evaluation metrics. Further, it facilitates effective debugging and error analysis of MT output via the use of dynamic filters that help users hone in on problem sentences with specific properties, such as genre, topic, sentence length, etc. The interface can be extended to include additional filters such as lexical, morphological, and syntactic features. Aside from helping debug MT output, it can also help in identifying problems in reference translations and evaluation metrics.
This paper proposes an efficient and semi-automated method for human-in-the-loop post- editing for machine translation (MT) corpus generation. The method is based on online training of a custom MT quality estimation metric on-the-fly as linguists perform post-edits. The online estimator is used to prioritize worse hypotheses for post-editing, and auto-close best hypothe- ses without post-editing. This way, significant improvements can be achieved in the resulting quality of post-edits at a lower cost due to reduced human involvement. The trained estimator can also provide an online sanity check mechanism for post-edits and remove the need for ad- ditional linguists to review them or work on the same hypotheses. In this paper, the effect of prioritizing with the proposed method on the resulting MT corpus quality is presented versus scheduling hypotheses randomly. As demonstrated by experiments, the proposed method im- proves the lifecycle of MT models by focusing the linguist effort on production samples and hypotheses, which matter most for expanding MT corpora to be used for re-training them
We present enhancements to a speech-to-speech translation pipeline in order to perform automatic dubbing. Our architecture features neural machine translation generating output of preferred length, prosodic alignment of the translation with the original speech segments, neural text-to-speech with fine tuning of the duration of each utterance, and, finally, audio rendering to enriches text-to-speech output with background noise and reverberation extracted from the original audio. We report and discuss results of a first subjective evaluation of automatic dubbing of excerpts of TED Talks from English into Italian, which measures the perceived naturalness of automatic dubbing and the relative importance of each proposed enhancement.
We describe a system to rapidly generate high-quality closed captions and subtitles for live broadcasted TV shows, using automated components, namely Automatic Speech Recognition and Machine Translation. The human stays in the loop for quality assurance and optional post-editing. We also describe how the system feeds the human edits and corrections back into the different components for improvement of these components and with that of the overall system. We finally describe the operation of this system in a real life environment within a broadcast network, where we implemented the system to transcribe, process broadcast transmissions and generate high-quality closed captions in Arabic and translate these into English subtitles in short time.
In this paper, we describe an extension to a hybrid machine translation system for handling dialect Arabic, using a decoding algorithm to normalize non-standard, spontaneous and dialectal Arabic into Modern Standard Arabic. We prove the feasibility of the approach by measuring and comparing machine translation results in terms of BLEU with and without the proposed approach. We show in our tests that on real-live broadcast input with transcriptions of dialectal speech we achieve an increase on BLEU of about 1%, and on web content with dialect text of about 2%.
CACI has developed and delivered systems for document exploitation and processing to Government customers around the world. Many of these systems include advanced language processing capabilities in order to enable rapid triage of vast collections of foreign language documents, separating the content that requires immediate human attention from the less immediately pressing material. AppTek provides key patent-pending Machine Translation technology for this critical process, rendering material in Arabic, Farsi and other languages into an English rendition that enables both further automated processing and rapid review by monolingual analysts, to identify the documents that require immediate linguist attention. Both CACI and AppTek have been working with customers to develop capabilities that enable them, the users, to be the ones in command of making their systems learn and continuously improve. We will describe how we put this critical user requirement into the systems and the key role that the user's involvement played in this. We will also discuss some of the key components of the system and what the customer-centric evolution of the system will be, including our document translation workflow, the machine translation technology within it, and our approaches to supporting the technology and sustaining its success designed around adapting to user needs.
In this paper, a system is presented that recognizes spoken utterances in Arabic Dialects which are translated into text in English. The input is recorded from a broadcast channel and recognized using automatic speech recognition that recognize Modern Standard Arabic and Iraqi Colloquial Arabic. The recognized utterances are normalized into Modern Standard Arabic and the output of this Modern Standard Arabic interlingua is then translated by a hybrid machine translation system, combining statistical and rule-based features.
In this paper, we describe some concepts of language models beyond the usually used standard trigram and use such language models for statistical machine translation. In statistical machine translation the language model is the a-priori knowledge source of the system about the target language. One important requirement for the language model is the correct word order, given a certain choice of words, and to score the translations generated by the translation model Pr(f1J/eI1), in view of the syntactic context. In addition to standard m-grams with long histories, we examine the use of Part-of-Speech based models as well as linguistically motivated grammars with stochastic parsing as a special type of language model. Translation results are given on the VERBMOBIL task, where translation is performed from German to English, with vocabulary sizes of 6500 and 4000 words, respectively.