Search

Popular musicians often learn music by ear. It is unclear what role technology plays for those with experience at this task. In search of opportunities for the development of novel human-recording interactions, we analyze 18 YouTube videos depicting real-world examples of by-ear learning, and discuss why, during this preliminary phase of research, online videos are appropriate data. From our observations we generate hypotheses that can inform future work. For example, a musician's scope of learning may influence what technological interactions would help them, they could benefit from tools that accommodate their working memory, and transcription does not appear to play a key role in ear learning. Based on these findings, we pose a number of research questions, and discuss their methodological considerations to guide future study.

Active Learning (AL) addresses the high costs of collecting human annotations by strategically annotating the most informative samples. However, for subjective NLP tasks, incorporating a wide range of perspectives in the annotation process is crucial to capture the variability in human judgments. We introduce Annotator-Centric Active Learning (ACAL), which incorporates an annotator selection strategy following data sampling. Our objective is two-fold: (1) to efficiently approximate the full diversity of human judgments, and (2) to assess model performance using annotator-centric metrics, which emphasize minority perspectives over a majority. We experiment with multiple annotator selection strategies across seven subjective NLP tasks, employing both traditional and novel, human-centered evaluation metrics. Our findings indicate that ACAL improves data efficiency and excels in annotator-centric performance evaluations. However, its success depends on the availability of a sufficiently large and diverse pool of annotators to sample from.

Large-scale survey tools enable the collection of citizen feedback in opinion corpora. Extracting the key arguments from a large and noisy set of opinions helps in understanding the opinions quickly and accurately. Fully automated methods can extract arguments but (1) require large labeled datasets that induce large annotation costs and (2) work well for known viewpoints, but not for novel points of view. We propose HyEnA, a hybrid (human + AI) method for extracting arguments from opinionated texts, combining the speed of automated processing with the understanding and reasoning capabilities of humans. We evaluate HyEnA on three citizen feedback corpora. We find that, on the one hand, HyEnA achieves higher coverage and precision than a state-of-the-art automated method when compared to a common set of diverse opinions, justifying the need for human insight. On the other hand, HyEnA requires less human effort and does not compromise quality compared to (fully manual) expert analysis, demonstrating the benefit of combining human and artificial intelligence., Comment: Published in JAIR

Understanding citizens' values in participatory systems is crucial for citizen-centric policy-making. We envision a hybrid participatory system where participants make choices and provide motivations for those choices, and AI agents estimate their value preferences by interacting with them. We focus on situations where a conflict is detected between participants' choices and motivations, and propose methods for estimating value preferences while addressing detected inconsistencies by interacting with the participants. We operationalize the philosophical stance that "valuing is deliberatively consequential." That is, if a participant's choice is based on a deliberation of value preferences, the value preferences can be observed in the motivation the participant provides for the choice. Thus, we propose and compare value estimation methods that prioritize the values estimated from motivations over the values estimated from choices alone. Then, we introduce a disambiguation strategy that addresses the detected inconsistencies between choices and motivations by directly interacting with the participants. We evaluate the proposed methods on a dataset of a large-scale survey on energy transition. The results show that explicitly addressing inconsistencies between choices and motivations improves the estimation of an individual's value preferences. The disambiguation strategy does not show substantial improvements when compared to similar baselines--however, we discuss how the novelty of the approach can open new research avenues and propose improvements to address the current limitations., Comment: Under review

Recent advances in NLP show that language models retain a discernible level of knowledge in deontological ethics and moral norms. However, existing works often treat morality as binary, ranging from right to wrong. This simplistic view does not capture the nuances of moral judgment. Pluralist moral philosophers argue that human morality can be deconstructed into a finite number of elements, respecting individual differences in moral judgment. In line with this view, we build a pluralist moral sentence embedding space via a state-of-the-art contrastive learning approach. We systematically investigate the embedding space by studying the emergence of relationships among moral elements, both quantitatively and qualitatively. Our results show that a pluralist approach to morality can be captured in an embedding space. However, moral pluralism is challenging to deduce via self-supervision alone and requires a supervised approach with human labels., Comment: To appear in Findings of EACL 2024

This extended abtract describes the preliminary qualitative results coming from a therapeutic laboratory focused on the use of the Pepper robot to promote autonomies and functional acquisitions in highly functioning (Asperger) children with autism. The field lab, ideated and led by a multidisciplinary team, involved 4 children, aged 11-13, who attended the laboratory sessions once a week for four months.

This paper introduces an experimental trial that will take place from February to June 2021, and which will see the use of the Pepper robot in a therapeutic laboratory on autonomies that will promote functional acquisitions in children diagnosed with high functioning autism/Asperger's syndrome., Comment: IUI Workshops 2021

Charge transport in organic semiconductors is notoriously extremely sensitive to the presence of disorder, both internal and external (i.e. related to the interactions with the dielectric layer), especially for n-type materials. Internal dynamic disorder stems from large thermal fluctuations both in intermolecular transfer integrals and (molecular) site energies in weakly interacting van der Waals solids and sources transient localization of the charge carriers. The molecular vibrations that drive transient localization typically operate at low-frequency (< a-few-hundred cm-1), which renders it difficult to assess them experimentally. Hitherto, this has prevented the identification of clear molecular design rules to control and reduce dynamic disorder. In addition, the disorder can also be external, being controlled by the gate insulator dielectric properties. Here we report on a comprehensive study of charge transport in two closely related n-type molecular organic semiconductors using a combination of temperature-dependent inelastic neutron scattering and photoelectron spectroscopy corroborated by electrical measurements, theory and simulations. We provide unambiguous evidence that ad hoc molecular design enables to free the electron charge carriers from both internal and external disorder to ultimately reach band-like electron transport., Comment: arXiv admin note: text overlap with arXiv:1909.05344

In the context of graphene-based composite applications, a complete understanding of charge conduction in multilayer reduced graphene oxides (rGO) is highly desirable. However, these rGO compounds are characterized by multiple and different sources of disorder depending on the chemical method used for their synthesis. Most importantly the precise role of interlayer interaction in promoting or jeopardizing electronic flow remains unclear. Here, thanks to the development of a multiscale computational approach combining first-principles calculations with large scale transport simulations, the transport scaling laws in multilayer rGO are unraveled, explaining why diffusion worsens with increasing film thickness. In contrast, contacted films are found to exhibit an opposite trend when the mean free path becomes shorter than the channel length, since conduction becomes predominantly driven by interlayer hopping. These predictions are favourably compared with experimental data and open a road towards the optimization of graphene-based composites with improved electrical conduction., Comment: 19 pages main text, 18 pages Supporting Information

Large area van der Waals (vdW) thin films are assembled materials consisting of a network of randomly stacked nanosheets. The multi-scale structure and the two-dimensional nature of the building block mean that interfaces naturally play a crucial role in the charge transport of such thin films. While single or few stacked nanosheets (i.e. vdW heterostructures) have been the subject of intensive works, little is known about how charges travel through multilayered, more disordered networks. Here we report a comprehensive study of a prototypical system given by networks of randomly stacked reduced graphene oxide 2D nanosheets, whose chemical and geometrical properties can be controlled independently, permitting to explore percolated networks ranging from a single nanosheet to some billions with room temperature resistivity spanning from 10-5 to 10-1 ohm m. We systematically observe a clear transition between two different regimes at a critical temperature T*: Efros-Shklovskii variable range hopping (ESVRH) below T* and power law (PL) behavior above. Firstly, we demonstrate that the two regimes are strongly correlated with each other, both depending on the charge localization length xi, calculated by ES-VRH model, which corresponds to the characteristic size of overlapping sp2 domains belonging to different nanosheets. Thus, we propose a microscopic model describing the charge transport as a geometrical phase transition, given by the metal-insulator transition associated with the percolation of quasi-1D nanofillers with length xi, showing that the charge transport behavior of the networks is valid for all geometries and defects of the nanosheets, ultimately suggesting a generalized description on vdW and disordered thin films.

We analyze the poset of moves in chip-firing, as defined by Klivans and Liscio. Answering a question of Propp, we show that the move poset forms the join-irreducibles of the poset of configurations. The proof involves a graph augmentation and an analysis of configurations in which only one firing move is available. We then use this framework to analyze the problem of chip-firing on a line, where the move poset is relevant to the problem of labeled chip-firing., Comment: 11 pages, 7 figures

Solid-state dewetting phenomenon in silver thin films offers a straightforward method to obtain structures having controlled shape or size -this latter in principle spanning several orders of magnitudes -- with potentially strong interest in many applications involving high-tech industry and biomedicine. In this work nanostructured silver is deposited by pulsed electron ablation technique and its surface modified upon thermal treatments in air at increasing temperatures. Surface chemistry and morphology are then monitored simultaneously by X-ray photoemission spectroscopy and atomic force microscopy; in particular, the power spectral density of surface heights is used to analyze the alteration of morphology induced by annealing. It is shown that this approach adds a level of information about the dewetting process since it allows to separate between long- and short-range surface behavior and to retrieve statistical quantities relevant to a description of the features in view of applications. Our results are presented in the framework of a multidisciplinary approach, advantages and limits of which are deepened and discussed.

In 2016, Hopkins, McConville, and Propp proved that labeled chip-firing on a line always leaves the chips in sorted order if the number of chips is even. We present a novel proof of this result. We then apply our methods to resolve a number of related conjectures concerning the confluence of labeled chip-firing systems., Comment: 23 pages, 9 figures, accepted as poster to FPSAC 2020, submitted to JCTA-A

Oligothiophenes are pi-conjugated semiconducting and fluorescent molecules whose self-assembly properties are widely investigated for application in organic electronics, optoelectronics, biophotonics and sensing. We report here an approach to the preparation of crystalline oli-gothiophene nano/micro-fibers based on the use of a sulfur overrich quaterthiophene building block, -T4S4-, containing in its covalent network all the information needed to promote the di-rectional, pi-pi stacking driven, self-assembly of Ar-T4S4-Ar oligomers into fibers with hierar-chical supramolecular arrangement from nano- to microscale. We show that when Ar varies from unsubstituted thiophene to thiophene substituted with electron withdrawing groups, a wide redistribution of the molecular electronic charge takes place without substantially affecting the aggregation modalities of the oligomer. In this way a structurally comparable series of fibers is obtained having progressively varying optical properties, redox potentials, photoconductivity and type of prevailing charge carriers (from p- to n-type). A thorough characterization of the fi-bers based on SEM, CD, CV, X-ray diffraction, UV-vis and PL spectroscopies, photoconductivi-ty and KPFM measurements is reported. With the aid of DFT calculations, combined with X-ray data, a model accounting for the growth of the fibers from molecular to nano- and microscale is proposed. We believe that the simple strategy outlined in this study allows to establish a straightforward correlation between the molecular structure of the components and the function-al properties of the corresponding self-assembled nano/micro-fibers.

The scarcity of Smart Home data is still a pretty big problem, and in a world where the size of a dataset can often make the difference between a poor performance and a good performance for problems related to machine learning projects, this needs to be resolved. But whereas the problem of retrieving real data can't really be resolved, as most of the time the process of installing sensors and retrieving data can be found to be really expensive and time-consuming, we need to find a faster and easier solution, which is where synthetic data comes in. Here we propose BinarySDG (Binary Synthetic Data Generator) as a flexible and easy way to generate synthetic data for binary sensors.

Charge transport in organic semiconductors is notoriously extremely sensitive to the presence of disorder, both intrinsic and extrinsic, especially for n-type materials. Intrinsic dynamic disorder stems from large thermal fluctuations both in intermolecular transfer integrals and (molecular) site energies in weakly interacting van der Waals solids and sources transient localization of the charge carriers. The molecular vibrations that drive transient localization typically operate at low-frequency (< a-few-hundred cm-1), which renders it difficult to assess them experimentally. Hitherto, this has prevented the identification of clear molecular design rules to control and reduce dynamic disorder. In addition, the disorder can also be extrinsic, being controlled by the gate insulator dielectric properties. Here we report on a comprehensive study of charge transport in two closely related n-type molecular organic semiconductors using a combination of temperature-dependent inelastic neutron scattering and photoelectron spectroscopy corroborated by electrical measurements, theory and simulations. We provide unambiguous evidence that ad hoc molecular design enables to free the electron charge carriers from both intrinsic and extrinsic disorder to ultimately reach band-like electron transport.

The chemical modification of electrodes with organic materials is a common approach to tune the electronic and electrostatic landscape between interlayers in optoelectronic devices, thus facilitating charge injection at the electrode/semiconductor interfaces and improving their performance. The use of photochromic molecules for the surface modification allows dynamic control of the electronic and electrostatic properties of the electrode and thereby enables additional functionalities in such devices. Here, we show that the electronic properties of a transparent indium tin oxide (ITO) electrode are reversibly and dynamically modified by depositing organic photochromic switches (diarylethenes) in the form of self-assembled monolayers (SAMs). By combining a range of surface characterization and density functional theory calculations, we present a detailed picture of the SAM binding onto ITO, the packing density of molecules, their orientation, as well as the work function modification of the ITO surface due to the SAM deposition. Upon illumination with ultraviolet and green light, we observe a reversible shift of the frontier occupied levels by 0.7 eV, and concomitantly a reversible work function change of ca. 60 meV. Our results prove the viability of dynamic switching of the electronic properties of the electrode with external light stimuli, which could be used to fabricate ITO-based photo-switchable optoelectronic devices.