[article-cris] Palvelut / Services

Permanent URI for this collection

https://aaltodoc.aalto.fi/handle/123456789/21536

Browse

Now showing 1 - 20 of 785

New data, benchmark and baseline for L2 speaking assessment for low-resource languages
(ISCA - International Speech Communication Association, 2023) Kurimo, Mikko; Getman, Yaroslav; Voskoboinik, Ekaterina; Al-Ghezi, Ragheb; Kallio, Heini; Kuronen, Mikko; von Zansen, Anna; Hilden, Raili; Kronholm, Sirkku; Huhta, Ari; Lindén, Krister; Speech Recognition; University of Jyväskylä; University of Helsinki; Department of Information and Communications Engineering
The development of large multilingual speech models provides the possibility to construct high-quality speech technology even for low-resource languages. In this paper, we present the speech data of L2 learners of Finnish and Finland Swedish that we have recently collected for training and evaluation of automatic speech recognition (ASR) and speaking assessment (ASA). It includes over 4000 recordings by over 300 students per language in short read-aloud and free-form tasks. The recordings have been manually transcribed and assessed for pronunciation, fluency, range, accuracy, task achievement, and a holistic proficiency level. We present also an ASR and ASA benchmarking setup we have constructed using this data and include results from our baseline systems built by fine-tuning self-supervised multilingual model for the target language. In addition to benchmarking, our baseline system can be used by L2 students and teachers for online self-training and evaluation of oral proficiency.
The Spores of Life and Death
(Aalto University School of Arts, Design and Architecture, 2023-09-29) Pusa, Tiina; Department of Art and Media; Department of Art and Media
Fungi are key players in ecosystems. They sustain life, affect the transformation of life forms, and are crucial actors in recycling carbon and minerals back into the common cycle. In addition to the author, three mushrooms are involved in the constellation of this article. In Western art history, self-portraits are intertwined with the theme of death in multiple ways. This posthumanist art-based study asks how mushrooms challenge our understanding of death. The study is located in the framework of queer death studies. Queering death by providing a basis for fungi to grow is a comforting thought.
We are all necronauts
(Aalto University School of Arts, Design and Architecture, 2023-09-29) Sederholm, Helena; Department of Art and Media; Department of Art and Media
In this thematic issue, we have a compilation of articles, visual essays, and a commentary dealing with death from diverse viewpoints. In many texts, posthuman and more-than-human-aspect, that is, the relation to other species, has been emphasized. Nevertheless, it is impossible to avoid the perspective of humans, especially when dealing with art, philosophy, and education.
Severity classification of Parkinson's disease from speech using single frequency filtering-based features
(International Speech Communication Association, 2023) Kadiri, Sudarsana; Kodali, Manila; Alku, Paavo; Speech Communication Technology; Department of Information and Communications Engineering
Developing objective methods for assessing the severity of Parkinson's disease (PD) is crucial for improving the diagnosis and treatment. This study proposes two sets of novel features derived from the single frequency filtering (SFF) method: (1) SFF cepstral coefficients (SFFCC) and (2) MFCCs from the SFF (MFCC-SFF) for the severity classification of PD. Prior studies have demonstrated that SFF offers greater spectrotemporal resolution compared to the short-time Fourier transform. The study uses the PC-GITA database, which includes speech of PD patients and healthy controls produced in three speaking tasks (vowels, sentences, text reading). Experiments using the SVM classifier revealed that the proposed features outperformed the conventional MFCCs in all three speaking tasks. The proposed SFFCC and MFCC-SFF features gave a relative improvement of 5.8% & 2.3% for the vowel task, 7.0% & 1.8% for the sentence task, and 2.4% & 1.1% for the read text task, in comparison to MFCC features.
Investigating wav2vec2 context representations and the effects of fine-tuning, a case-study of a Finnish model
(International Speech Communication Association, 2023-08-20) Grosz, Tamas; Getman, Yaroslav; Al-Ghezi, Ragheb; Rouhe, Aku; Kurimo, Mikko; Speech Recognition; Department of Information and Communications Engineering; Department of Information and Communications Engineering
Self-supervised speech models, such as the wav2vec2, have become extremely popular in the past few years. Their main appeal is that after their pre-training on a large amount of audio, they require only a small amount of supervised, finetuning data to achieve outstanding results. Despite their immense success, very little is understood about the pre-trained models and how finetuning changes them. In this work, we take the first steps towards a better understanding of wav2vec2 systems using model interpretation tools such as visualization and latent embedding clustering. Through our analysis, we gain new insights into the abilities of the pre-trained networks and the effect that finetuning has on them. We demonstrate that the clusters learned by the pre-trained model are just as important a factor as the supervised training data distribution in determining the accuracy of the finetuned system, which could aid us in selecting the most suitable pre-trained model for the supervised data.
Spectral warping based data augmentation for low resource children’s speaker verification
(Springer, 2023-11-03) Kathania, Hemant Kumar; Kadyan, Virender; Kadiri, Sudarsana Reddy; Kurimo, Mikko; Department of Information and Communications Engineering; University of Petroleum and Energy Studies; Department of Information and Communications Engineering
In this paper, we present our effort to develop an automatic speaker verification (ASV) system for low resources children’s data. For the children’s speakers, very limited amount of speech data is available in majority of the languages for training the ASV system. Developing an ASV system under low resource conditions is a very challenging problem. To develop the robust baseline system, we merged out of domain adults’ data with children’s data to train the ASV system and tested with children’s speech. This kind of system leads to acoustic mismatches between training and testing data. To overcome this issue, we have proposed spectral warping based data augmentation. We modified adult speech data using spectral warping method (to simulate like children’s speech) and added it to the training data to overcome data scarcity and mismatch between adults’ and children’s speech. The proposed data augmentation gives 20.46% and 52.52% relative improvement (in equal error rate) for Indian Punjabi and British English speech databases, respectively. We compared our proposed method with well known data augmentation methods: SpecAugment, speed perturbation (SP) and vocal tract length perturbation (VTLP), and found that the proposed method performed best. The proposed spectral warping method is publicly available at https://github.com/kathania/Speaker-Verification-spectral-warping .
HENGITYSTIEAEROSOLIPÄÄSTÖJEN MITTAAMINEN JA HUOMIOINTI TARTUNTATAUTIEN EHKÄISYSSÄ
(2023) Saari, Sampo; Tuhkuri Matvejeff, Anna; Silvonen, Ville; Heikkilä, Paavo; Hakala, Jani; Sanmark, Enni; Oksanen, Lotta-Maria; Rönkkö, Topi; Taipale, Aimo; Laukkanen, Anne-Maria; Alku, Paavo; Geneid, Ahmed; Tampere University of Applied Sciences; University of Helsinki; Tampere University; VTT Technical Research Centre of Finland Ltd.; Informaatio- ja tietoliikennetekniikan laitos
Patogeenien aerosolileviäminen on monimutkainen tapahtuma, jossa on paljon tekijöitä, jotka vaikuttavat infektioriskiin. Yksi tärkeä muuttuja on hengitysteistä vapautuvien aerosolihiukkasten määrä ja ominaisuudet. Tässä esityksessä keskitytään hengitysteistä syntyvien aerosolihiukkaspäästön mekanismeihin, dynamiikkaan, mittauksiin, alustaviin tuloksiin sekä niiden huomiointiin tartuntatautien ehkäisyssä. Kokeellista tutkimusta varten kehitettiin uusi siirrettävä mittausjärjestelmä, jolla pystyttiin mittaamaan koehenkilöiden hengitystieaerosoleja reaaliaikaisesti laajalla hiukkaskokoalueella. Laitteiston avulla pystyttiin mittaamaan absoluuttisia hiukkaspäästökertoimia koehenkilöille erilaisissa harjoitustilanteissa (mm. puhe, laulu, kuiskaus, yskiminen).
Dynamics of convivial affective atmospheres
(Elsevier, 2023-07) Rokka, Joonas; Auriacombe, Brigitte; Arnould, Eric; Sitz, Lionel; EMLYON Business School; Department of Marketing; Department of Marketing
This article examines the dynamics of affective atmospheres in explaining compelling touristic service experiences. Extending affect-theoretical research, we theorize and examine the role of affective bodily encounters through which convivial atmospheres are (re)produced. Based on multi-sited ethnographic research in iconic Club Med resorts, our findings emphasize the spontaneity and fragility of convivial affective atmospheres, meaning they are difficult to control, but also how their ongoing “bubbling” is shaped by materiality, ritual activities, and temporality. We contribute by (1) developing the concept of convivial affective atmosphere, and by (2) offering a theoretical framework that helps advance research into the affective entanglement and dynamics of atmospheres in tourism. Finally, we critically discuss what service providers can do to facilitate and re-produce convivial atmospheres.
Kaksi toksisuustestiä syötävien ja myrkyllisten sienten sekä sisätilapölyjen analytiikassa
(2021-03-09) Hintikka, Tuomas; Andersson, Maria; Andersson, Magnus; Lundell, Taina; Kurnitski, Jarek; Salonen, Heidi; University of Helsinki; Rakennetekniikka, mekaniikka ja laskenta; Ahola, Mervi; Merikari, Anna
Sisäilma-analytiikassa terveyshaitan arvioiminen solutoksisuustesteillä on haasteellista, koska altistusaika, haitta-aineiden alkuperä ja aika altistuksen sekä oireiden ilmenemisen välillä ovat usein tuntemattomia. Vertasimme suomalaisten metsäsienten raportoitua in vivo myrkyllisyyttä in vitro ja ex vivo -solutoksisuustestien vasteisiin. Osoitimme että 89 % tunnetuista myrkyllisistä sienilajeista (n=189 ja 19 % tunnetuista syötävistä sienistä (n=26) olivat solutestissä toksisia. Kosteista "hometaloista" kerätyistä pölyistä (n=13) 46 % ja kuivien talojen pölyistä (n=14) 57 % olivat toksisia. Rakennusten pölyjen ominaistoksisuus voi viitata haitalliseen altistukseen, mutta ei kerro onko toksinen aine mikrobi- tai ihmisperäinen. Vasta toksisuustestien ja kemiallisen analytiikan yhdistäminen mahdollistaisi myrkyllisten aineiden ja päästölähteiden tunnistamisen.
The Future of Enterprise Information Systems
(Springer, 2023-10-30) Sunyaev, Ali; Dehling, Tobias; Strahringer, Susanne; Da Xu, Li; Heinig, Martin; Perscheid, Michael; Alt, Rainer; Rossi, Matti; Karlsruhe Institute of Technology; Technische Universität Dresden; Old Dominion University; SAP Research; University of Potsdam; Leipzig University; Department of Information and Service Management; Department of Information and Service Management
NOPEA JA HALPA MENETELMÄ STACHYBOTRYS-KASVUN TOTEAMISEEN SISÄTILOISSA - KOLME ESIMERKKIÄ
(2023) Andersson, Aino; Rakennusten suunnittelu, rakentaminen ja toimivuus
Stachybotrys home vaatii kasvaakseen korkean kosteuden, sen itiöt ovat isoja, 5 μm ×10 μm, leviävät huonosti ilmateitse ja ovat helppoja tunnistaa valomikroskoopissa. Jos Stachybotrys itiöita löytyy runsaasti laskeutuneesta sisätilapölystä, ne tuskin ovat peräisin ulkoilmasta, eli niiden läsnäolo saattaa viitata kasvustoon rakennuksen sisällä. Tutkimuksessa kehitettiin menetelmä, jolla todennäköisen Stachybotrys kasvuston voi löytää sisätiloista. Tässä julkaisussa kuvataan kolme esimerkkitapausta, joissa menetelmän avulla Stachybotrys-kasvusto pystyttiin paikantamaan. Menetelmä perustuu itiöiden rikastamiseen kyllästyneen NaCl-liuoksen avulla laskeutuneesta pölystä, ja niiden tunnistamiseen. Erilaisia menetelmiä voidaan hyödyntää sisäilmaongelmien havaitsemisessa ja mahdollisten sisäilmaongelmien syiden selvittämisessä.
Joint mixed-timescale optimization of content caching and delivery policy in NOMA-based vehicular networks
(Elsevier, 2023-12) Liu, Jingyao; Feng, Guangsheng; Zhang, Zhibo; Zheng, Liying; Wang, Huiqiang; Hämäläinen, Jyri; Harbin Engineering University; Department of Information and Communications Engineering; Department of Information and Communications Engineering
Recently, the development of Internet of Vehicles (IoV) and the increasing popularity of video applications have led to the fast-growing in-car video demand causing numerous challenges in wireless networks. Pre-caching and non-orthogonal multiple access (NOMA) have been regarded as two effective techniques to alleviate the mentioned challenge. In this paper, we propose a cache-aided cooperative transmission to maximize the quality of service (QoS) in the NOMA-based vehicular network. A QoS-oriented joint optimization problem is formulated, which incorporates power allocation, content caching, and delivery strategy. Considering, on the one hand, the slow update rate of cache content and, on the other hand, frequent handovers of vehicles between different transmitters, a mixed-timescale optimization is proposed where the serving cache is updated in a long-term phase, while content delivery and power allocation are optimized in a short-term phase. In the proposed approach, content caching is determined based on future user requests, vehicle tracking, and other delivery information. To make this possible, we leverage a substantial number of stochastic samples to approximate content caching in the long-term caching phase. Due to the NOMA-based transmission and integral variables, the setting leads to a Mixed Integer Non-Linear Programming (MINLP) problem, which is NP-hard. To solve this problem, an iterative method based on sample average approximation (SAA) and Successive Convex Approximation (SCA) is applied. Simulations demonstrate that the proposed algorithm can achieve better QoS than other recently proposed transmission schemes.
Data-Oriented Downlink RSMA Systems
(IEEE, 2023-10-01) Can, Mehmet; Ilter, Mehmet C.; Altunbas, Ibrahim; Istanbul Technical University; Risto Wichman Group; Department of Information and Communications Engineering
Rate-splitting multiple access (RSMA) provides a flexible and promising non-orthogonal multiple access paradigm which relies on splitting user messages into common and private parts and utilizing successive interference cancellation at the receivers. By doing so, the achievable degrees of freedom can be increased so that different frameworks are available, ranging from the infinite blocklength regime to the finite blocklength regime. The latter one attracted considerable attention over the fifth generation (5G) and beyond 5G networks in the context of short packet transmission. In this respect, data-oriented approach introduces a transient performance metric for small data transmissions where the amount of data and available bandwidth play an essential role in the performance evaluation. Motivated by this fact, this letter represents the first framework where a data-oriented approach is applied to downlink RSMA systems under finite blocklength regime. Particularly, the optimization of the RSMA downlink systemin the context of the data-oriented approach is proposed and the numerical results show that data-oriented RSMA systems introduce an efficient design for multi-user short-packet transmissions.
SecrecyPerformance Analysis of Backscatter Communications with Side Information
(MDPI AG, 2023-10-10) Kaveh, Masoud; Rostami Ghadi, Farshad; Jäntti, Riku; Yan, Zheng; Network Security and Trust; University of Málaga; Communication Engineering; Xidian University; Department of Information and Communications Engineering
Backscatter communication (BC) systems are a promising technology for internet of things (IoT) applications that allow devices to transmit information by modulating ambient radio signals without the need for a dedicated power source. However, the security of BC systems is a critical concern due to the vulnerability of the wireless channel. This paper investigates the impact of side information (SI) on the secrecy performance of BC systems. SI mainly refers to the additional knowledge that is available to the communicating parties beyond transmitted data, which can be used to enhance reliability, efficiency, security, and quality of service in various communication systems. In particular, in this paper, by considering a non-causally known SI at the transmitter, we derive compact analytical expressions of average secrecy capacity (ASC) and secrecy outage probability (SOP) for the proposed system model to analyze how SI affects the secrecy performance of BC systems. Moreover, a Monte Carlo simulation validates the accuracy of our analytical results and reveals that considering such knowledge at the transmitter has constructive effects on the system performance and ensures reliable communication with higher rates than the conventional BC systems without SI, namely, lower SOP and higher ASC are achievable.
TOIMISTOTILOJEN LÄMPÖOLOSUHTEET MUUTTUVAILMAVIRTAJÄRJESTELMÄLLÄ KÄYTETTÄESSÄ ERITYYPPISIÄ TULOILMALAITTEITA
(2023) Mustakallio, Panu; Wu, Yuxin; Kosonen, Risto; Kaukola, Tuomas; Energiakonversio ja järjestelmät; Chongqing University; Halton Oy
Muuttuvailmavirtajärjestelmää käytettäessä toimistotilojen ilmanvaihtoa säädetään tyypillisesti käyttöasteen ja lämpökuormien mukaan. Sisäilmaston hallinnan tavoitteena on ylläpitää hyvät lämpöolosuhteet ja tuloilmanjako eri tilanteissa. Tässä tutkimuksessa tutkittiin tarpeenmukaisen ilmanvaihdon toimintaa eri kattohajottajien tapauksissa toimistotiloissa. Tutkimus tehtiin täyden mittakaavan kokeilla ja CFD-simuloinneilla. Lämpöolosuhteet mitattiin kalibroiduilla ilman nopeus-, turbulenssi- ja lämpötila-antureilla. Näiden kautta laskettiin paikallinen vetoriski oleskeluvyöhykkeellä. Tutkimusten tuloksena havaittiin aktiivisen radiaalihajottajan tapauksissa tasaisimmat lämpöolosuhteet. Pienillä ilmavirroilla tuloilmasuihkun irtoamisesta aiheutuvaa liiallista vetoriskiä oleskeluvyöhykkeellä ei havaittu minkään kattohajottajan tapauksessa
OVERHEATING RISK OF APARTMENTS IN THE HELSINKI REGION DURING THE HOT SUMMER OF 2021
(2023) Velashjerdi Farahani, Azin; Jokisalo, Juha; Korhonen, Natalia; Jylhä, Kirsti; Kosonen, Risto; Energiakonversio ja järjestelmät; Finnish Meteorological Institute
During heat waves, room air temperatures without cooling can raise health risk levels. Therefore, it is important to study the actual room air temperatures in residential buildings during hot summers. In this study, the hourly indoor temperatures of 6057 apartments during the hot summer of 2021 were measured and analyzed in the Helsinki region. These apartments are different in size and age. The results showed that the indoor temperature in 96% of the apartments exceeded 27 C,, 34% exceeded 30 C, and around 5% of them exceeded 32 C. The results showed a significantly high risk of overheating in most of the apartments despite differences in room number and age.
Dataset and analysis of acoustic intensity vector measurements around an upscaled ear model
(2023) Geldert, Aaron; Marschall, Marton; Pulkki, Ville; Communication Acoustics: Spatial Sound and Psychoacoustics; Department of Information and Communications Engineering
This work presents a dataset of acoustic intensity vector measurements taken around a 3D-printed, upscaled model of an ear. Impulse responses of the acoustic vector were collected in an anechoic chamber in a dense grid of measurement positions around the pinna. Automated interleaved swept-sine measurements were carried out using an arc of loudspeakers, a robotic mover, a probe consisting of a pressure sensor and three anemometer particle velocity sensors, with the ear mounted on a turntable. In total, 145 unique directions of arrival and 225 measurement points were considered. Results show that the measured data is suitable for high resolution time, frequency, and spatial analyses. The full dataset containing impulse responses, ear model geometry, calibration data, and example processing scripts is shared in an open-access repository. Potential applications include the analysis and visualization of pinna-related transfer functions, as well as the validation of numerical acoustical modeling techniques.
Think like a robot : How interactions with humanoid service robots affect consumers’ decision strategies
(Elsevier Ltd, 2023-10-03) Liao, Jiancai; Huang, Jingya; Department of Marketing; Antai College of Economics and Management; Department of Marketing
Service robots are beginning to be a popular choice in the frontline of customer service. It is important to explore how interacting with humanoid service robots (vs. human employees) influences consumers' following behaviors. Specifically, the present research examines how interacting with humanoid service robots (vs. human employees) influences consumers' emotional intensity and subsequent decision strategies. Through four experiments, the current research demonstrates that interacting with humanoid service robots (vs. human employees) induces a lower emotional intensity, which in turn promotes a relative reliance on cognitive reasoning (vs. affective feeling) in consumers’ decision-making such as hotel room choice. Furthermore, this research identifies the moderating role of service provider gender and anthropomorphism tendency. The main effect disappears when the service robot is female (vs. male) and when consumers have a higher anthropomorphism tendency. The present research provides significant implications for both robot literature and the tourism and hospitality business.
Representing Inner Voices in Virtual Reality Environments
(2023) Parkkola, Kuura; McKenzie, Thomas; Häkkinen, Jukka; Pulkki, Ville; Dept Signal Process and Acoust; University of Helsinki; Communication Acoustics: Spatial Sound and Psychoacoustics; Department of Information and Communications Engineering
The inner auditory experience comprises various sounds which, rather than originating from sources in their environment, form as a result of internal processes within the brain of an observer. Examples of such sounds are, for instance, verbal thoughts and auditory hallucinations. Traditional audiovisual media representations of inner voices have tended to focus on impact and storytelling, rather than aiming to reflect a true-to-life experience. In virtual reality (VR) environments, where plausibility is favoured over this hyper-real sound design, a question remains on the best ways to recreate realistic, and on the other hand, entertaining inner and imagined voices via head-tracked headphones and spatial audio tools. This paper first presents a questionnaire which has been completed by 70 participants on their own experience of inner voices. Next, the results of the questionnaire are used to inform a VR experiment, whereby different methods to render inner voices are compared. This is conducted using a short film created for this project. Results show that people mostly expect realism from the rendering of inner voices and auditory hallucinations when the focus is on believability. People’s expectations for inner voice did not change considerably in an entertainment context, whereas for hallucinations, exaggerated reverberation was preferred.
Location as Supervision for Weakly Supervised Multi-Channel Source Separation of Machine Sounds
(2023) Falcon-Perez, Ricardo; Wichern, Gordon; Germain, Francois G.; Le Roux, Jonathan; Communication Acoustics: Spatial Sound and Psychoacoustics; Department of Information and Communications Engineering
In this work, we are interested in learning a model to separate sources that cannot be recorded in isolation, such as parts of a machine that must run simultaneously in order for the machine to function. We assume the presence of a microphone array and knowledge of the source locations (potentially obtained from schematics or an auxiliary sensor such as a camera). Our method uses the source locations as weak labels for learning to separate the sources, since we cannot obtain the isolated source signals typically used as training targets. We propose a loss function that requires the directional features computed from the separated sources to match the true direction of arrival for each source, and also include a reconstruction loss to ensure all frequencies are taken into account by at least one of the separated sources output by our model. We benchmark the performance of our algorithm using synthetic mixtures created using machine sounds from the DCASE 2021 Task 2 dataset in challenging reverberant conditions. While reaching lower objective scores than a model with access to isolated source signals for training, our proposed weakly-supervised model obtains promising results and applies to industrial scenarios where collecting isolated source signals is prohibitively expensive or impossible.

Browse

Recent Submissions