A low-latency graph computer to identify metastable particles at the Large Hadron Collider for real-time analysis of potential dark matter signatures.

Image recognition is a pervasive task in many information-processing environments. We present a solution to a difficult pattern recognition problem that lies at the heart of experimental particle physics. Future experiments with very high-intensity beams will produce a spray of thousands of particles in each beam-target or beam-beam collision. Recognizing the trajectories of these particles as they traverse layers of electronic sensors is a massive image recognition task that has never been accomplished in real time. We present a real-time processing solution that is implemented in a commercial field-programmable gate array using high-level synthesis. It is an unsupervised learning algorithm that uses techniques of graph computing. A prime application is the low-latency analysis of dark-matter signatures involving metastable charged particles that manifest as disappearing tracks.

Have you been there before? Decoding recognition of spatial scenes from fMRI signals in precuneus.

One potential application of forensic "brain reading" is to test whether a suspect has previously experienced a crime scene. Here, we investigated whether it is possible to decode real life autobiographic exposure to spatial locations using fMRI. In the first session, participants visited four out of eight possible rooms on a university campus. During a subsequent scanning session, subjects passively viewed pictures and videos from these eight possible rooms (four old, four novel) without giving any responses. A multivariate searchlight analysis was employed that trained a classifier to distinguish between "seen" versus "unseen" stimuli from a subset of six rooms. We found that bilateral precuneus encoded information that can be used to distinguish between previously seen and unseen rooms and that also generalized to the two stimuli left out from training. We conclude that activity in bilateral precuneus is associated with the memory of previously visited rooms, irrespective of the identity of the room, thus supporting a parietal contribution to episodic memory for spatial locations. Importantly, we could decode whether a room was visited in real life without the need of explicit judgments about the rooms. This suggests that recognition is an automatic response that can be decoded from fMRI data, thus potentially supporting forensic applications of concealed information tests for crime scene recognition.

Evolution and emerging trends of named entity recognition: Bibliometric analysis from 2000 to 2023.

Identifying valuable information within the extensive texts documented in natural language presents a significant challenge in various disciplines. Named Entity Recognition (NER), as one of the critical technologies in text data processing and mining, has become a current research hotspot. To accurately and objectively review the progress in NER, this paper employs bibliometric methods. It analyzes 1300 documents related to NER obtained from the Web of Science database using CiteSpace software. Firstly, statistical analysis is performed on the literature and journals that were obtained to explore the distribution characteristics of the literature. Secondly, the core authors in the field of NER, the development of the technology in different countries, and the leading institutions are explored by analyzing the number of publications and the cooperation network graph. Finally, explore the research frontiers, development tracks, research hotspots, and other information in this field from a scientific point of view, and further discuss the five research frontiers and seven research hotspots in depth. This paper explores the progress of NER research from both macro and micro perspectives. It aims to assist researchers in quickly grasping relevant information and offers constructive ideas and suggestions to promote the development of NER.

Putative Risk Biomarkers of Bipolar Disorder in At-risk Youth.

Bipolar disorder is a highly heritable and functionally impairing disease. The recognition and intervention of BD especially that characterized by early onset remains challenging. Risk biomarkers for predicting BD transition among at-risk youth may improve disease prognosis. We reviewed the more recent clinical studies to find possible pre-diagnostic biomarkers in youth at familial or (and) clinical risk of BD. Here we found that putative biomarkers for predicting conversion to BD include findings from multiple sample sources based on different hypotheses. Putative risk biomarkers shown by perspective studies are higher bipolar polygenetic risk scores, epigenetic alterations, elevated immune parameters, front-limbic system deficits, and brain circuit dysfunction associated with emotion and reward processing. Future studies need to enhance machine learning integration, make clinical detection methods more objective, and improve the quality of cohort studies.

Getting to Know Named Entity Recognition: Better Information Retrieval.

Named entity recognition (NER) is a powerful computer system that utilizes various computing strategies to extract information from raw text input, since the early 1990s. With rapid advancement in AI and computing, NER models have gained significant attention and been serving as foundational tools across numerus professional domains to organize unstructured data for research and practical applications. This is particularly evident in the medical and healthcare fields, where NER models are essential in efficiently extract critical information from complex documents that are challenging for manual review. Despite its successes, NER present limitations in fully comprehending natural language nuances. However, the development of more advanced and user-friendly models promises to improve work experiences of professional users significantly.

Neural basis of language familiarity effects on voice recognition: An fNIRS study.

Recognizing talkers' identity via speech is an important social skill in interpersonal interaction. Behavioral evidence has shown that listeners can identify better the voices of their native language than those of a non-native language, which is known as the language familiarity effect (LFE). However, its underlying neural mechanisms remain unclear. This study therefore investigated how the LFE occurs at the neural level by employing functional near-infrared spectroscopy (fNIRS). Late unbalanced bilinguals were first asked to learn to associate strangers' voices with their identities and then tested for recognizing the talkers' identities based on their voices speaking a language either highly familiar (i.e., native language Chinese), or moderately familiar (i.e., second language English), or completely unfamiliar (i.e., Ewe) to participants. Participants identified talkers the most accurately in Chinese and the least accurately in Ewe. Talker identification was quicker in Chinese than in English and Ewe but reaction time did not differ between the two non-native languages. At the neural level, recognizing voices speaking Chinese relative to English/Ewe produced less activity in the inferior frontal gyrus, precentral/postcentral gyrus, supramarginal gyrus, and superior temporal sulcus/gyrus while no difference was found between English and Ewe, indicating facilitation of voice identification by the automatic phonological encoding in the native language. These findings shed new light on the interrelations between language ability and voice recognition, revealing that the brain activation pattern of the LFE depends on the automaticity of language processing.

No, Bionic Reading does not work.

It has recently been claimed that presenting text with the first half of each word printed in bold (as is done in this example), so-called Bionic Reading, facilitates reading. However, empirical tests of this claim are lacking, and theoretically one might expect a cost rather than a benefit. Here I tested participants' reading speed of 100 paragraphs that were presented either in 'Bionic' or in normal font. Statistical analyses revealed no significant difference in reading times between Bionic and normal reading. I conclude that Bionic Reading does not facilitate reading.

Multi-scene application of intelligent inspection robot based on computer vision in power plant.

As industries develop, the automation and intelligence level of power plants is constantly improving, and the application of patrol robots is also increasingly widespread. This research combines computer vision technology and particle swarm optimization algorithm to build an obstacle recognition model and obstacle avoidance model of an intelligent patrol robot in a power plant respectively. Firstly, the traditional convolutional recurrent neural network is optimized, and the obstacle recognition model of an intelligent patrol robot is built by combining the connection timing classification algorithm. Then, the artificial potential field method optimizes the traditional particle swarm optimization algorithm, and an obstacle avoidance model of an intelligent patrol robot is built. The performance of the two models was tested, and it was found that the highest precision, recall, and F1 values of the identification model were 0.978, 0.974, and 0.975. The highest precision, recall, and F1 values of the obstacle avoidance model were 0.97, 0.96, and 0.96 respectively. The two optimization models designed in this research have better performance. In conclusion, the two models in this study are superior to the traditional methods in recognition effect and obstacle avoidance efficiency, providing an effective technical scheme for intelligent patrol inspection of power plants.

Facing up to Problem Gambling: Tracing the Emergence of Facial Recognition Technology as a means of Enforcing Voluntary Self-Exclusion.

Computer technology has long been touted as a means of increasing the effectiveness of voluntary self-exclusion schemes - especially in terms of relieving gaming venue staff of the task of manually identifying and verifying the status of new customers. This paper reports on the government-led implementation of facial recognition technology as part of an automated self-exclusion program in the city of Adelaide in South Australia-one of the first jurisdiction-wide enforcements of this controversial technology in small venue gambling. Drawing on stakeholder interviews, site visits and documentary analysis over a two year period, the paper contrasts initial claims that facial recognition offered a straightforward and benign improvement to the efficiency of the city's long-running self-excluded gambler program, with subsequent concerns that the new technology was associated with heightened inconsistencies, inefficiencies and uncertainties. As such, the paper contends that regardless of the enthusiasms of government, tech industry and gaming lobby, facial recognition does not offer a ready 'technical fix' to problem gambling. The South Australian case illustrates how this technology does not appear to better address the core issues underpinning problem gambling, and/or substantially improve conditions for problem gamblers to refrain from gambling. As such, it is concluded that the gambling sector needs to pay close attention to the practical outcomes arising from initial cases such as this, and resist industry pressures for the wider replication of this technology in other jurisdictions.

Information perseveration in recognition memory: Examining the scope of sequential dependencies.

Models of recognition memory often assume that decisions are made independently from each other. Yet there is growing evidence that consecutive recognition responses show sequential dependencies, whereby making one response increases the probability of repeating that response from one trial to the next trial. Across six experiments, we replicated this response-related carryover effect using word and nonword stimuli and further demonstrated that the content of the previous trial-both perceptual and conceptual-can also bias the response to the current test probe, with both perceptual (orthographic) and conceptual (semantic) similarity boosting the probability of consecutive "old" responses. Finally, a manipulation of attentional engagement in Experiments 3a and 3b provided little evidence these carryover effects on recognition decisions are merely a product of lapses in attention. Taken together, the current study reinforces prior findings that recognition decisions are not made independently, and that multiple forms of information perseverate across consecutive trials.

Visual word identification beyond common words: The role of font and letter case in brand names.

While abstractionist theories of visual word recognition propose that perceptual elements like font and letter case are filtered out during lexical access, instance-based theories allow for the possibility that these surface details influence this process. To disentangle these accounts, we focused on brand names embedded in logotypes. The consistent visual presentation of brand names may render them much more susceptible to perceptual factors than common words. In the present study, we compared original and modified brand logos, varying in font or letter case. In Experiment 1, participants decided whether the stimuli corresponded to existing brand names or not, regardless of graphical information. In Experiment 2, participants had to categorize existing brand names semantically - whether they corresponded to a brand in the transportation sector or not. Both experiments showed longer response times for the modified brand names, regardless of font or letter-case changes. These findings challenge the notion that only abstract units drive visual word recognition. Instead, they favor those models that assume that, under some circumstances, the traces in lexical memory may contain surface perceptual information.

Machine Learning Assisted Self-Powered Identity Recognition Based on Thermogalvanic Hydrogel for Intelligent Security.

Identity recognition as the first barrier of intelligent security plays a vital role, which is facing new challenges that are unable to meet the need of intelligent era due to low accuracy, complex configuration and dependence on power supply. Here, a finger temperature-driven intelligent identity recognition strategy is presented based on a thermogalvanic hydrogel (TGH) by actively discerning biometric characteristics of fingers. The TGH is a dual network PVA/Agar hydrogel in an H2O/glycerol binary solvent with [Fe(CN)6]3-/4- as a redox couple. Using a concave-arranged TGH array, the characteristics of users can be distinguished adequately even under an open environment by extracting self-existent intrinsic temperature features from five typical sites of fingers. Combined with machine learning, the TGH array can recognize different users with a high average accuracy of 97.6%. This self-powered identity recognition strategy is further applied to a smart lock, attaining a more reliable security protection from biometric characteristics than bare passwords. This work provides a promising solution for achieving better identity recognition, which has great advantages in intelligent security and human-machine interaction toward future Internet of everything.

The application of graphic language personalized emotion in graphic design.

Emotion Recognition is the experience of attitude in graphic language expression and composition. People use both verbal and non-verbal behaviours to communicate their emotions. Visual communication and graphic design are always evolving to meet the demands of an increasingly affluent and culturally conscious populace. When graphic designing works, designers should consider their own opinions about related works from the audience's or customer's standpoint so that the emotion between them can resonate. Hence, this study proposes a personalized emotion recognition framework based on convolutional neural networks (PERF-CNN) to create visual content for graphic design. Graphic designers prioritize the logic of showing objects in interactive designs and use visual hierarchy and page layout approaches to respond to users' demands via typography and imagery. This ensures that the user experience is maximized. This research identifies three tiers of emotional thinking: expressive signal, emotional experience, and emotional infiltration, all of which affect graphic design. This article explores the subject of graphic design language and its ways of emotional recognition, as well as the relationship between graphic images, shapes, and feelings. CNN initially extracted expressive features from the user's face images and the poster's visual information. The clustering process categorizes the poster or advertisement images into positive, negative, and neutral classes. Research and applications of graphic design language benefit from the proposed method's experimental results, demonstrating that it outperforms conventional classification approaches in the dataset. In comparison to other popular models, the experimental results demonstrate that the proposed PERF-CNN model improves each of the following: classification accuracy (97.4 %), interaction ratio (95.6 %), emotion recognition ratio (98.9 %), rate of influence of pattern and colour features (94.4 %), and prediction error rate (6.5 %).

Audio-Based Emotion Recognition Using Self-Supervised Learning on an Engineered Feature Space.

Emotion recognition models using audio input data can enable the development of interactive systems with applications in mental healthcare, marketing, gaming, and social media analysis. While the field of affective computing using audio data is rich, a major barrier to achieve consistently high-performance models is the paucity of available training labels. Self-supervised learning (SSL) is a family of methods which can learn despite a scarcity of supervised labels by predicting properties of the data itself. To understand the utility of self-supervised learning for audio-based emotion recognition, we have applied self-supervised learning pre-training to the classification of emotions from the CMU Multimodal Opinion Sentiment and Emotion Intensity (CMU- MOSEI)'s acoustic data. Unlike prior papers that have experimented with raw acoustic data, our technique has been applied to encoded acoustic data with 74 parameters of distinctive audio features at discrete timesteps. Our model is first pre-trained to uncover the randomly masked timestamps of the acoustic data. The pre-trained model is then fine-tuned using a small sample of annotated data. The performance of the final model is then evaluated via overall mean absolute error (MAE), mean absolute error (MAE) per emotion, overall four-class accuracy, and four-class accuracy per emotion. These metrics are compared against a baseline deep learning model with an identical backbone architecture. We find that self-supervised learning consistently improves the performance of the model across all metrics, especially when the number of annotated data points in the fine-tuning step is small. Furthermore, we quantify the behaviors of the self-supervised model and its convergence as the amount of annotated data increases. This work characterizes the utility of self-supervised learning for affective computing, demonstrating that self-supervised learning is most useful when the number of training examples is small and that the effect is most pronounced for emotions which are easier to classify such as happy, sad, and angry. This work further demonstrates that self-supervised learning still improves performance when applied to the embedded feature representations rather than the traditional approach of pre-training on the raw input space.

Comparison of Labsan Tricell-1000 and Dirui FUS-2000 automated urine analyzers with manual microscopy.

Objectives: Urinalysis is a first-line test for screening for urinary tract infection. Several devices performing strip and sediment analysis have been introduced. The aim of this study was to compare the performance of Labsan Tricell-1000 and Dirui FUS-2000 automated urine analyzers with manual microscopy. Methods: 463 urine samples were analyzed. Digital image processing and particle recognition automatically display the cells in a flowing sheath fluid mixed monolayer urine sample, take the pictures of particles via digital camera, analyse these pics with a particle recognition software, transfer images of the formed elements to the screen and allow well-trained personnel to select, reclassify or remove them. Manual microscopy was used for comparison. Results: Agreement between Tricell-100 and manual microscopy was very good for RBC (Ï° = 0.80), and WBC (Ï° = 0.83); good for CaOx (Ï° = 0.69), SEC (Ï° = 0.80), YLC (Ï° = 0.72), HC (0.69) and LC (Ï° = 0.64); moderate for BAC (Ï° = 0.51), APC (Ï° = 0.43) and MT (Ï° = 0.55); fair for GC (Ï° = 0.39) and RTEC (Ï° = 0.32). Conclusions: Labsan Trion TriCell-1000 demonstrated satisfactory performance and can be used in routine urinalysis. In the case of low counts of RBC, presence of yeast, crystal, casts or cell clumping in urine sediment, characterization of urine particles should be performed by manual microscopy.

Expandable-RCNN: toward high-efficiency incremental few-shot object detection.

This study aims at addressing the challenging incremental few-shot object detection (iFSOD) problem toward online adaptive detection. iFSOD targets to learn novel categories in a sequential manner, and eventually, the detection is performed on all learned categories. Moreover, only a few training samples are available for all sequential novel classes in these situations. In this study, we propose an efficient yet suitably simple framework, Expandable-RCNN, as a solution for the iFSOD problem, which allows online sequentially adding new classes with zero retraining of the base network. We achieve this by adapting the Faster R-CNN to the few-shot learning scenario with two elegant components to effectively address the overfitting and category bias. First, an IOU-aware weight imprinting strategy is proposed to directly determine the classifier weights for incremental novel classes and the background class, which is with zero training to avoid the notorious overfitting issue in few-shot learning. Second, since the above zero-retraining imprinting approach may lead to undesired category bias in the classifier, we develop a bias correction module for iFSOD, named the group soft-max layer (GSL), that efficiently calibrates the biased prediction of the imprinted classifier to organically improve classification performance for the few-shot classes, preventing catastrophic forgetting. Extensive experiments on MS-COCO show that our method can significantly outperform the state-of-the-art method ONCE by 5.9 points in commonly encountered few-shot classes.

siRNA that participates in Drosophila dosage compensation is produced by many 1.688X and 359 bp repeats.

Organisms with differentiated sex chromosomes must accommodate unequal gene dosage in males and females. Male fruit flies increase X-linked gene expression to compensate for hemizygosity of their single X chromosome. Full compensation requires localization of the Male-Specific Lethal (MSL) complex to active genes on the male X, where it modulates chromatin to elevate expression. The mechanisms that identify X chromatin are poorly understood. The euchromatic X is enriched for AT-rich, ∼359 bp satellites termed the 1.688X repeats. Autosomal insertions of 1.688X DNA enable MSL recruitment to nearby genes. Ectopic expression of dsRNA from one of these repeats produces siRNA and partially restores X-localization of MSLs in males with defective X recognition. Surprisingly, expression of double stranded RNA from three other 1.688X repeats failed to rescue males. We reconstructed dsRNA-expressing transgenes with sequence from two of these repeats and identified phasing of repeat DNA, rather than sequence or orientation, as the factor that determines rescue of males with defective X recognition. Small RNA sequencing revealed that siRNA was produced in flies with a transgene that rescues, but not in those carrying a transgene with the same repeat but different phasing. We demonstrate that pericentromeric X heterochromatin promotes X-recognition through a maternal effect, potentially mediated by small RNA from closely related heterochromatic repeats. This suggests that the sources of siRNAs promoting X recognition are highly redundant. We propose that enrichment of satellite repeats on Drosophilid X chromosomes facilitates the rapid evolution of differentiated sex chromosomes by marking the X for compensation.

Streamlining social media information extraction for public health research with deep learning.

OBJECTIVE: Social media-based public health research is crucial for epidemic surveillance, but most studies identify relevant corpora with keyword-matching. This study develops a system to streamline the process of curating colloquial medical dictionaries. We demonstrate the pipeline by curating a UMLS-colloquial symptom dictionary from COVID-19-related tweets as proof of concept. METHODS: COVID-19-related tweets from February 1, 2020, to April 30, 2022 were used. The pipeline includes three modules: a named entity recognition module to detect symptoms in tweets; an entity normalization module to aggregate detected entities; and a mapping module that iteratively maps entities to Unified Medical Language System concepts. A random 500 entity sample were drawn from the final dictionary for accuracy validation. Additionally, we conducted a symptom frequency distribution analysis to compare our dictionary to a pre-defined lexicon from previous research. RESULTS: We identified 498,480 unique symptom entity expressions from the tweets. Pre-processing reduces the number to 18,226. The final dictionary contains 38,175 unique expressions of symptoms that can be mapped to 966 UMLS concepts (accuracy = 95%). Symptom distribution analysis found that our dictionary detects more symptoms and is effective at identifying psychiatric disorders like anxiety and depression, often missed by pre-defined lexicons. CONCLUSIONS: This study advances public health research by implementing a novel, systematic pipeline for curating symptom lexicons from social media data. The final lexicon's high accuracy, validated by medical professionals, underscores the potential of this methodology to reliably interpret and categorize vast amounts of unstructured social media data into actionable medical insights across diverse linguistic and regional landscapes.

Attentional decoder networks for chest X-ray image recognition on high-resolution features.

BACKGROUND AND OBJECTIVE: This paper introduces an encoder-decoder-based attentional decoder network to recognize small-size lesions in chest X-ray images. In the encoder-only network, small-size lesions disappear during the down-sampling steps or are indistinguishable in the low-resolution feature maps. To address these issues, the proposed network processes images in the encoder-decoder architecture similar to U-Net families and classifies lesions by globally pooling high-resolution feature maps. However, two challenging obstacles prohibit U-Net families from being extended to classification: (1) the up-sampling procedure consumes considerable resources, and (2) there needs to be an effective pooling approach for the high-resolution feature maps. METHODS: Therefore, the proposed network employs a lightweight attentional decoder and harmonic magnitude transform. The attentional decoder up-samples the given features with the low-resolution features as the key and value while the high-resolution features as the query. Since multi-scaled features interact, up-sampled features embody global context at a high resolution, maintaining pathological locality. In addition, harmonic magnitude transform is devised for pooling high-resolution feature maps in the frequency domain. We borrow the shift theorem of the Fourier transform to preserve the translation invariant property and further reduce the parameters of the pooling layer by an efficient embedding strategy. RESULTS: The proposed network achieves state-of-the-art classification performance on the three public chest X-ray datasets, such as NIH, CheXpert, and MIMIC-CXR. CONCLUSIONS: In conclusion, the proposed efficient encoder-decoder network recognizes small-size lesions well in chest X-ray images by efficiently up-sampling feature maps through an attentional decoder and processing high-resolution feature maps with harmonic magnitude transform. We open-source our implementation at https://github.com/Lab-LVM/ADNet.