This paper presents XLS-R, a large-scale model for cross-lingual speech representation learning based on wav2vec 2.0. We train models with up to 2B parameters on nearly half a million hours of publicly available speech audio in 128 languages, an order of magnitude more public data than the largest known prior work. Our evaluation covers a wide range of tasks, domains, data regimes and languages, both high and low-resource. On the CoVoST-2 speech translation benchmark, we improve the previous state of the art by an average of 7.4 BLEU over 21 translation directions into English. For speech recognition, XLS-R improves over the best known prior work on BABEL, MLS, CommonVoice as well as VoxPopuli, lowering error rates by 14-34% relative on average. XLS-R also sets a new state of the art on VoxLin-gua107 language identification. Moreover, we show that with sufficient model size, cross-lingual pretraining can perform as well as English-only pretraining when translating English speech into other languages, a setting which favors monolingual pretraining. We hope XLS-R can help to improve speech processing tasks for many more languages of the world. Models and code are available at www.github.com/ pytorch/fairseq/tree/master/examples/wav2vec/xlsr. 1 * Equal contribution. † Work done while at Facebook AI. ‡ Equal advising.

While the general idea of self-supervised learning is identical across modalities, the actual algorithms and objectives differ widely because they were developed with a single modality in mind. To get us closer to general self-supervised learning, we present data2vec, a framework that uses the same learning method for either speech, NLP or computer vision. The core idea is to predict latent representations of the full input data based on a masked view of the input in a selfdistillation setup using a standard Transformer architecture. Instead of predicting modality-specific targets such as words, visual tokens or units of human speech which are local in nature, data2vec predicts contextualized latent representations that contain information from the entire input. Experiments on the major benchmarks of speech recognition, image classification, and natural language understanding demonstrate a new state of the art or competitive performance to predominant approaches. Models and code are available at www.github.com/pytorch/fairseq/ tree/master/examples/data2vec.

Most purulent orofacial infections are of odontogenic origin. It is well established that odontogenic infections are polymicrobial in nature. Empiric antibiotics were administered before the culture and sensitivity test results were obtained and specific antibiotics were administered based on the culture and sensitivity test results. But resistance was a challenging problem all throughout along with development of more virulent strains of microorganisms which were more infectious and resistant to many known antibiotics. Objective: To identify the causative aerobic and anaerobic microorganisms responsible for orofacial infections and to evaluate the resistance against empirical antibiotics used in the treatment of space infections. Method: 142 patients with head and neck fascial space infections of odontogenic origin were randomly taken, the pus samples and aspirates were collected aseptically from patients for aerobic and anaerobic microbiological study. Results: In this study the most common aerobic organism isolated was streptococcus viridians (34.49%), most common anaerobe was peptostreptococci, (61.11%) and the most common mixed organism was streptococcus with peptostreptococci (30%). Amoxicillin was the most commonly used empirical drug in all cases and showed highest resistance (96.55%) for all the organisms. But linezolid (100%) was sensitive to all the aerobic, anaerobic and mixed group of organisms. Metronidazole (100%) turned out to be sensitive to the entire anaerobic group. Clindamycin (100%) appeared sensitive to the entire aerobic group. Conclusion: Knowledge about the pathologic flora involved in head and neck infection in a locality and their sensitivity and resistance to commonly used antibiotics will help the clinician in administering appropriate antibiotics.

Fourteen sugarcane genotypes were screened for traits contributing to high water use efficiency (WUE) and temperature tolerance in a field experiment by imposing moisture stress at formative phase i.e. 40 DAP to 120 DAP (days after planting). WUE was measured using surrogate methods viz., SPAD chlorophyll meter readings (SCMR), specific leaf area (SLA) and thermostability was quantified by measuring membrane relative injury percentage on 20 days (60 DAP) and 80 days (120 DAP) after imposing moisture stress (DAIS). Among the fourteen genotypes, four genotypes viz., 97 R 129, 92 V 104, CO (O) 061 and CO 6907 showed low SLA and higher SCMR values, which indicates higher water use efficiency. However, higher thermostability was recorded in 92 V 104 and CO 6907genotypes only. Hence, the 92 V 104 and CO 6907 genotypes having both WUE and temperature tolerance traits can be used as donar parents in breeding programmes. SCMR, SLA, relarive injury per cent can be used as surrogate methods for measurement of WUE and temperature tolerance in sugarcane and can be used as selection traits in breeding programme for developing drought tolerant genotypes.

No abstract