Are Word Suggestions Beneficial? The Effect of Typing Efficiency and Suggestion Accuracy

The conventional P300-based character spelling BCI system consists of a character presentation paradigm and a classification system. In this paper, we propose modifications to both in order to increase the word typing speed and accuracy. In the paradigm part, we have modified the T9 (Text on Nine keys) interface which is similar to the keypad of mobile phones being used for text messaging. Then we have integrated a custom-built dictionary to give word suggestions to a user while typing. The user can select one out of the given suggestions to complete word typing. Our proposed paradigms significantly reduce the word typing time and make words typing more convenient by typing complete words with only few initial character spellings. In the classification part we have adopted a Random Forest (RF) classifier. The RF improves classification accuracy by combining multiple decision trees. We conducted experiments with five subjects using the proposed BCI system. Our results demonstrate that our system increases typing speed significantly: our proposed system took an average time of 1.83 minutes per word, while typing ten random words, whereas the conventional spelling required 3.35 minutes for the same words under the same conditions, decreasing the typing time by 45.37%.

Suggesting words to complete a given sequence of characters is a common feature of typing interfaces. Yet, previous studies have not found a clear benefit, some even finding it detrimental. We report on the first study to control for two important factors, word suggestion accuracy and typing efficiency. Our accuracy factor is enabled by a new methodology that builds on standard metrics of word suggestions. Typing efficiency is based on device type. Results show word suggestions are used less often in a desktop condition, with little difference between tablet and phone conditions. Very accurate suggestions do not improve entry speed on desktop, but do on tablet and phone. Based on our findings, we discuss implications for the design of automation features in typing systems.

PressTapFlick: Exploring a gaze and foot-based multimodal approach to gaze typing

With the rapid development of the Internet of Things (IoT) technology, the number and types of devices accessing the Internet are increasing, leading to increased network security problems such as hacker attacks and botnets. Usually, these attacks are related to the type of device, and the risk can be effectively reduced if the type of network device can be efficiently identified and controlled. The traditional network device identification method uses active detection technology to obtain information about the device and match it with a manually defined fingerprint database to achieve network device identification. This method impacts the smoothness of the network and requires the manual establishment of fingerprint libraries, which imposes a large labor cost but only achieves a low identification efficiency. The traditional machine learning method only considers the information of individual packets; it does not consider the timing relationship between packets, and the recognition effect is poor. Based on the above research, in this paper, we considered the packet temporal relationship, proposed the TCN model of the Inception structure, extracted the packet temporal relationship, and designed a multi-head self-attention mechanism to fuse the features to generate device fingerprints for device identification. Experiments were conducted on the publicly available UNSW dataset, and the results showed that this method achieved notable improvements compared to the traditional machine learning method, with F1 reaching 96.76%.

The complex interplay between gaze tracker accuracy and interface design is the focus of this paper. Two slightly different variants of GazeTalk, a hierarchical typing interface, were contrasted with a novel interface, Dasher, in which text entry is done by continuous navigation. All of the interfaces were tested with a good and a deliberate bad calibration of the tracker. The purpose was to investigate, if performance indices normally used for evaluation of typing systems, such as characters per minute (CPM) and error-rate, could differentiate between the conditions, and thus guide an iterative system development of both trackers and interfaces. Gaze typing with one version of the static, hierarchical menu systems was slightly faster than the others. Error measures, in terms of rate of backspacing, were also significantly different for the systems, while the deliberate bad tracker calibrations did not have any measurable effect. Learning effects were evident under all conditions. Power-law-of-practice learning models suggested that Dasher might be more efficient than GazeTalk in the long run.

The conventional P300 BCI system for character spelling is typically composed of a paradigm that displays flashing characters and a classifier which identifies target characters. Typically a user has to type each character of a word at a time: this spelling process is slow and it can take several minutes to type an entire word. In this work, we propose a new word typing scheme by integrating a word suggestion mechanism via a dictionary search into the conventional P300-based speller. Our new P300-based word typing system consists of an initial character spelling paradigm, a smart dictionary unit to give suggestions of possible words, and the final word selection paradigm to select a word out of the suggestions. Our proposed methodology reduces typing time significantly and makes word typing more convenient. We have tested our system with four subjects and our results demonstrate an average words typing time of 1.66 minute, whereas the conventional took 2.9 minute for the same words.

A P300-based brain computer interface system for words typing

The RSVP Keyboard™ is a brain-computer interface (BCI)-based typing system for people with severe physical disabilities, specifically those with locked-in syndrome (LIS). It uses signals from an electroencephalogram (EEG) combined with information from an n-gram language model to select letters to be typed. One characteristic of the system as currently configured is that it does not keep track of past EEG observations, i.e., observations of user intent made while the user was in a different part of a typed message. We present a principled approach for taking all past observations into account, and show that this method results in a 20% increase in simulated typing speed under a variety of conditions on realistic stimuli. We also show that this method allows for a principled and improved estimate of the probability of the backspace symbol, by which mis-typed symbols are corrected. Finally, we demonstrate the utility of automatically typing likely letters in certain contexts, a technique that achieves increased typing speed under our new method, though not under the baseline approach.

\n An important issue with eye-based typing iis the correct identification of both whrn the userselects a key and which key is selected. Traditional solutions are based on predefined gaze fixation time, known as dwell-time methods. In an attempt to improve accuracy long dwell times are adopted, which un turn lead to fatigue and longer response limes. These problems motivate the proposal of methods free of dwell-time, or with very short ones, which rely on more robust recognition techniques to reduce the uncertainty about user\\'s actions. These techniques are specially important when the users have disabilities which affect their eye movements or use inexpensive eye trackers. An approach to deal with the recognition problem is to treat it as a spelling correction task. An usual strategy for spelling correction is to model the problem as the transmission of a word through a noisy-channel, such that it is necessary to determine which known word of a lexicon is the received string. A feasible application of this method requires the reduction of the set of candidate words by choosing only the ones that can be transformed into the imput by applying up to k character edit operations. This idea works well on traditional typing because the number of errors per word is very small. However, this is not the case for eye-based typing systems, which are much noiser. In such a scenario, spelling correction strategies do not scale well as they grow exponentially with k and the lexicon size. Moreover, the error distribution in eye typing is different, with much more insertion errors due to specific sources, of noise such as the eye tracker device, particular user behaviors, and intrinsic chracteeristics of eye movements. Also, the lack of a large corpus of errors makes it hard to adopt probabilistic approaches based on information extracted from real world data. To address all these problems, we propose an effective recognition approach by combining estimates extracted from general error corpora with domain-specific knowledge about eye-based input. The technique is ablçe to calculate edit disyances effectively by using a Mor-Fraenkel index, searchable using a minimun prfect hashing. The method allows the early processing of most promising candidates, such that fast pruned searches present negligible loss in word ranking quality. We also propose a linear heuristic for estimating edit-based distances which take advantage of information already provided by the index. Finally, we extend our recognition model to include the variability of the eye movements as source of errors, provide a comprehensive study about the importance of the noise model when combined with a language model and determine how it affects the user behaviour while she is typing. As result, we obtain a method very effective on the task of recognizing words and fast enough to be use in real eye typing systems. In a transcription experiment with 8 users, they archived 17.46 words per minute using proposed model, a gain of 11.3% over a state-of-the-art eye-typing system. The method was particularly userful in more noisier situations, such as the first use sessions. Despite significant gains in typing speed and word recognition ability, we were not able to find statistically significant differences on the participants\\' perception about their expeience with both methods. This indicates that an improved suggestion ranking may not be clearly perceptible by the users even when it enhances their performance.\n

Text entry is a critical capability for any modern computing experience, with lightweight augmented reality (AR) glasses being no exception. Designed for all-day wearability, a limitation of lightweight AR glass is the restriction to the inclusion of multiple cameras for extensive field of view in hand tracking. This constraint underscores the need for an additional input device. We propose a system to address this gap: a ring-based mid-air gesture typing technique, RingGesture, utilizing electrodes to mark the start and end of gesture trajectories and inertial measurement units (IMU) sensors for hand tracking. This method offers an intuitive experience similar to raycast-based mid-air gesture typing found in VR headsets, allowing for a seamless translation of hand movements into cursor navigation. To enhance both accuracy and input speed, we propose a novel deep-learning word prediction framework, Score Fusion, comprised of three key components: a) a word-gesture decoding model, b) a spatial spelling correction model, and c) a lightweight contextual language model. In contrast, this framework fuses the scores from the three models to predict the most likely words with higher precision. We conduct comparative and longitudinal studies to demonstrate two key findings: firstly, the overall effectiveness of RingGesture, which achieves an average text entry speed of 27.3 words per minute (WPM) and a peak performance of 47.9 WPM. Secondly, we highlight the superior performance of the Score Fusion framework, which offers a 28.2% improvement in uncorrected Character Error Rate over a conventional word prediction framework, Naive Correction, leading to a 55.2% improvement in text entry speed for RingGesture. Additionally, RingGesture received a System Usability Score of 83 signifying its excellent usability.

Designing a comfortable, familiar, and efficient one-handed text entry method for Head-Mounted Displays (HMDs) remains a significant challenge. Existing midair typing systems induce fatigue, while novel techniques often demand extensive training or sacrifice input efficiency. Consequently, we introduce ThumbSwype , a novel thumb-to-finger text entry technique that adapts smartphone swipe typing for HMDs. Users see the traditional QWERTY keyboard overlaid on their index, middle, and ring fingers, allowing them to perform swipe gestures with their thumb to type words. In an evaluation study (N=16) , participants achieved a mean of 14.52 words per minute (WPM), which is 63.8% of their smartphone swipe-typing performance, with a peak average of 20.2 WPM. We compare ThumbSwype’s performance with related work, and discuss directions for future improvement.

Since the report of a molecular thin-film organic solar cell (OSC) by Tang, organic materials have increasingly become attractive candidates for the fabrication of cost-efficient and flexible photovoltaic cells. In particular, polymer bulkheterojunction (BHJ) solar cells based on interpenetrating networks of an electron donor and an acceptor, with a largearea donor and acceptor interface, resulting in an efficient photo-induced charge separation, have gained considerable interest. Despite of their relatively low efficiency in comparison to conventional inorganic solar cells, the potential of roll-to-roll processing on low-cost and flexible substrates makes polymer solar cells (PSCs) so attractive as a cost-effective solution to the energy problems we are facing today. Among the various BHJ systems, poly(3-hexylthiophene) (P3HT) and [6,6]-phenyl-C61butyric acid methyl ester (PCBM) networks produced by spin-coating a blend combined with a preor post-heat treatment to improve the degree of ordering of P3HT have shown the highest efficiencies up to ca. 4–5 % under 80 or 100 mW cm illumination under AM 1.5 G condition. Because the electrical and optical properties of conjugated polymers are strongly dependent on the structural order of polymers, the processing methods and conditions, which determine the degree of organization of polymers, have a critical influence on the performance of electrical and optoelectrical devices based on these polymers. In addition, considering efficiencies of ca. 5 %, the development of novel solution processes that are compatible with low-cost mass production is one of the crucial requirements for practical device applications. Here, we report on novel solution-processed high-efficiency polymer solar cells based on a blend of P3HT and PCBM, resulting from improved organization of the P3HT. By directly brush painting a blend of P3HT and PCBM on appropriately temperature-controlled substrates, enhanced ordering of the polymers, induced by the shear stress in the direction parallel to the brushing direction, was achieved. This highly ordered active layer facilitates charge transport separated at the interface of the P3HT and PCBM, leading to an increased efficiency, in particularly an improved fill factor. In addition, this novel solution process can be considered a promising method for the fabrication of flexible and large-area polymer solar cells based on high-throughput roll-to-roll manufacturing, which would make the realization of low-cost PSCs possible. First, two types of devices were fabricated by conventional spin-coating and brush painting, respectively, without any preor post-heat treatment, which are usually performed to stabilize a nanoscale interpenetrating network with a crystalline order, resulting in an increase in overall conversion efficiency. Figure 1 shows the schematic of the brushing method and the resulting I–V curves for the two types of devices. The brush painting was performed on a poly(3,4-ethylenedioxythiophene) (PEDOT):poly(styrene sulfonate) (PSS)-coated indium tin oxide (ITO) substrate on a hot plate with a temperature of 50 °C. A general paintbrush made of nylon fibrils was used and the active layer was coated with a speed of ca. 1.5 cm s. It took ca. 2 s to prepare a complete film on the ITO substrate with a size of 1.5 cm × 1.5 cm by brush painting twice. During the brush painting process, an appropriate temperature was necessary to make a smooth and uniform active layer with a high degree of ordering, which is related to the evaporation rate of the solvents. In our case chlorobenzene was used, for which good-quality films could not be obtained when the brush painting was carried out below 50 °C because of too slow evaporation of the solvent at low temperatures. To prepare active layers with the same thickness, the blend was spin-coated at room temperature on the PEDOT:PSScoated ITO substrate at 2000 rpm and the thickness was determined to be ca. 90 nm by means of a surface profiler (Kosaka ET-3000i). The surface morphologies determined by atomic force microscopy (AFM) for the surfaces produced by the brushing technique (with a rms roughness of ca. 0.91 nm) were similar to that of a conventional spin-coated active layer (not shown here). As shown in Figure 1b, the performance was improved when the active layer was prepared by the brush-painting process. In the case of a spin-coated device, a poor performance was observed with VOC (opencircuit voltage) = 0.6 V, ISC (short-circuit current density) = 3.59 mA cm, FF (fill factor) = 32.5 %, and ge (power conversion efficiency) = 0.7 %, which is similar to previous C O M M U N IC A TI O N

Emotion theories are mainly classified as categorical or dimensional approaches. Given the importance of emotional words in emotion research, researchers have constructed a co-occurrence corpus of 7 types of emotion words through word co-occurrence and big data corpora. However, in addition to the categorical approach, the dimensional approach plays an important role in natural language processing. In particular, valence has an important influence on the study of emotion and language. In this study, the co-occurrence corpus of 7 types of emotion words constructed by Chen et al. [1] was expanded to create a corpus of emotional valences. Then, stepwise multiple regression analysis was performed with the predicted criterion variables and 15 predictor variables. The criterion variables were the emotional valences of 553 frequently occurring stimulus words included in the Chinese Word Association Norms [2]. The predictor variables included the emotion co-occurrences scores for 2 clusters (a cluster of literal emotion words and a cluster of metaphorical emotion words) and 7 types of emotions (happiness, love, surprise, sadness, anger, disgust, and fear) [the emotional words were common words from both the co-occurrence corpus of 7 types of emotion words constructed by Chen et al. [1] and the Chinese Word Association Norms established by Hu et al. [2]] and the virtue word co-occurrences score. The results showed that the scores for literal happiness word co-occurrences, metaphorical happiness word co-occurrences, literal disgust word co-occurrences, literal fear word co-occurrences, and virtue word co-occurrences could predict the valence values of emotion words, with the multiple correlation coefficients of multiple regression analyses reaching .729. Subsequently, the valence values of 33,669 words were established using the formula obtained from the multiple regression analysis of the 553 words. Next, the correlation between the actual valence values and the predicted valence values was analyzed to test the cross-validity of the established valences using the common words in the norm established by Lee and Lee [3] for the emotionality ratings and free associations of 267 common Chinese words. The results showed that the correlation between the 2 was .755, indicating that the predicted values generated by the big data corpora and word co-occurrence had a degree of similarity with the manually determined values. Based on theories and tests, this study used the co-occurrence data of 7 emotions and virtue to construct the corpus of emotional valences for 33,669 Chinese words. The results showed that the combined use of big data corpora and word co-occurrence can effectively expand existing corpora that were established based on emotional categories, improve the efficiency of manual construction of corpora, and establish a larger corpus of emotional words. KeywordsEmotionValenceWord co-occurrenceBig dataChinese

People with motor impairments often have difficulty typing using desktop keyboards. We developed TrueKeys, a system that combines models of word frequency, keyboard layout, and typing error patterns to automatically identify and correct typing mistakes. In this paper, we describe the TrueKeys algorithm, compare its performance to existing correction algorithms, and report on a study of TrueKeys with 9 motor-impaired and 9 non-impaired participants. Running in non-interactive mode, TrueKeys performed more corrections than popular commercial and open source spell checkers. Used interactively, both motor-impaired and non-impaired users performed typing tasks significantly more accurately with TrueKeys than without. However, typing speed was reduced while TrueKeys was enabled.

PURPOSE: Young adults have the capacity to manage dual task conditions with minimal impairment to either the cognitive or the motor task; however, this ability decreases with age. Previous research on dual tasking has primarily examined over ground walking and minimal information is available on dual tasking on a treadmill. The purpose of this study was to examine the effect of dual tasking using a treadmill desk on changes in cognitive performance and gait parameters in young adults (YA) and middle-aged adults (MA). METHODS: YA (n=24; mean age 21.1±1.6 yrs) and MA (n=25; mean age 53.0±5.3 yrs) were recruited to participate in this study. Participants completed five cognitive tests (Stroop Word Color Test, phoneme monitoring, typing test, Sternberg working memory test, and serial 7 subtractions) in a single task (ST) and dual task (DT) condition in a randomized and counterbalanced order. Participants were seated at a desk for ST and walked on a treadmill desk at self-selected speed (mean speed YA=1.5±0.4 mph; MA=1.4±0.5 mph) during DT. An OptoGait system recorded gait parameters of step length, stride length, and coefficient of variation. RESULTS: There were no significant differences in gait parameters or test scores in YA when comparing DT and ST conditions (p>0.05). MA performed worse on word recall score (89.7±11.3 vs 96.6±7.5%, p=0.03), typing speed (44.9±11.2 vs 49.9±13.3 wpm, p=0.00), and Sternberg reaction time (1.5±2.0 vs 1.0±1.9 s, p=0.00) during the DT compared to the ST condition. MA stride length decreased during DT in the Sternberg test (37.7±5.9 vs 36.5±5.3 in, p=0.01) and serial 7 subtractions (37.2±5.7 vs 36.5±5.3 in, p=0.00). MA showed detriments in reaction times on the Stroop test (0.8±1.2 vs 0.6±0.9 s, p=0.00) and Sternberg test (1.5±2.0 vs 0.9±1.5 s, p=0.02) and decreased word recall score (89.7±11.3 vs 97.5±7.0%, p=0.04) compared to YA during the DT condition. CONCLUSION: The impairments in gait and cognitive test scores in MA but not YA suggest that the ability to simultaneously process cognitive demands and treadmill walking requirements decreases with age. Using a treadmill desk might affect work-related performance or gait parameters in middle-aged adults.