Tallinn University of Technology

About: Tallinn University of Technology is a education organization based out in Tallinn, Estonia. It is known for research contribution in the topics: European union & Computer science. The organization has 3688 authors who have published 10313 publications receiving 145058 citations. The organization is also known as: Tallinn Technical University & Tallinna Tehnikaülikool.

Temporary business model innovation – SMEs’ innovation response to the Covid-19 crisis

Abstract: The Covid‐19 crisis has hit SMEs particularly hard. Numerous business models (BM) have been limited or rendered downright impossible due to decreased social contact. SMEs can respond to this exogenous crisis via temporary business model innovation (BMI). This empirical study investigates these temporary BMs using a multiple case study approach based on five SMEs in Austria, Germany, and Liechtenstein who within a short period of time applied their core competencies and networks to integrate new BMs, which were in some cases very different from existing ones. These had a positive effect on strategic flexibility, and if desired can also be incorporated into the firm long‐term. The paper contributes to SME crisis management during the Covid‐19 pandemic by pointing out and developing a successful management mechanism that allows to survive a crisis or even improve during this time. Moreover, we contribute to BMI literature by explaining temporary BMI as a new form of BMI. It also makes clear to managers that temporary BMs add value to firms and create new revenue streams.

VoxLingua107: a Dataset for Spoken Language Recognition

Abstract: This paper investigates the use of automatically collected web audio data for the task of spoken language recognition. We generate semi-random search phrases from language-specific Wikipedia data that are then used to retrieve videos from YouTube for 107 languages. Speech activity detection and speaker diarization are used to extract segments from the videos that contain speech. Post-filtering is used to remove segments from the database that are likely not in the given language, increasing the proportion of correctly labeled segments to 98%, based on crowd-sourced verification. The size of the resulting training set (VoxLingua107) is 6628 hours (62 hours per language on the average) and it is accompanied by an evaluation set of 1609 verified utterances. We use the data to build language recognition models for several spoken language identification tasks. Experiments show that using the automatically retrieved training data gives competitive results to using hand-labeled proprietary datasets. The dataset is publicly available.

Multi-omics approach to study the growth efficiency and amino acid metabolism in Lactococcus lactis at various specific growth rates

Abstract: Lactococcus lactis is recognised as a safe (GRAS) microorganism and has hence gained interest in numerous biotechnological approaches. As it is fastidious for several amino acids, optimization of processes which involve this organism requires a thorough understanding of its metabolic regulations during multisubstrate growth. Using glucose limited continuous cultivations, specific growth rate dependent metabolism of L. lactis including utilization of amino acids was studied based on extracellular metabolome, global transcriptome and proteome analysis. A new growth medium was designed with reduced amino acid concentrations to increase precision of measurements of consumption of amino acids. Consumption patterns were calculated for all 20 amino acids and measured carbon balance showed good fit of the data at all growth rates studied. It was observed that metabolism of L. lactis became more efficient with rising specific growth rate in the range 0.10 - 0.60 h-1, indicated by 30% increase in biomass yield based on glucose consumption, 50% increase in efficiency of nitrogen use for biomass synthesis, and 40% reduction in energy spilling. The latter was realized by decrease in the overall product formation and higher efficiency of incorporation of amino acids into biomass. L. lactis global transcriptome and proteome profiles showed good correlation supporting the general idea of transcription level control of bacterial metabolism, but the data indicated that substrate transport systems together with lower part of glycolysis in L. lactis were presumably under allosteric control. The current study demonstrates advantages of the usage of strictly controlled continuous cultivation methods combined with multi-omics approach for quantitative understanding of amino acid and energy metabolism of L. lactis which is a valuable new knowledge for development of balanced growth media, gene manipulations for desired product formation etc. Moreover, collected dataset is an excellent input for developing metabolic models.

Long-haul optical transmission link using low-noise phase-sensitive amplifiers

Abstract: The capacity and reach of long-haul fiber optical communication systems is limited by in-line amplifier noise and fiber nonlinearities. Phase-sensitive amplifiers add 6 dB less noise than conventional phase-insensitive amplifiers, such as erbium-doped fiber amplifiers, and they can provide nonlinearity mitigation after each span. Realizing a long-haul transmission link with in-line phase-sensitive amplifiers providing simultaneous low-noise amplification and nonlinearity mitigation is challenging and to date no such transmission link has been demonstrated. Here, we demonstrate a multi-channel-compatible and modulation-format-independent long-haul transmission link with in-line phase-sensitive amplifiers. Compared to a link amplified by conventional erbium-doped fiber amplifiers, we demonstrate a reach improvement of 5.6 times at optimal launch powers with the phase-sensitively amplified link operating at a total accumulated nonlinear phase shift of 6.2 rad. The phase-sensitively amplified link transmits two data-carrying waves, thus occupying twice the bandwidth and propagating twice the total power compared to the phase-insensitively amplified link.