Broad Range Of Tasks Research Articles

AttacKG+: Boosting attack graph construction with Large Language Models

Attack graph construction seeks to convert textual cyber threat intelligence (CTI) reports into structured representations, portraying the evolutionary traces of cyber attacks. Even though previous research has proposed various methods to construct attack graphs, they generally suffer from limited generalization capability to diverse knowledge types as well as requirement of expertise in model design and tuning. Addressing these limitations, we seek to utilize Large Language Models (LLMs), which have achieved enormous success in a broad range of tasks given exceptional capabilities in both language understanding and zero-shot task fulfillment. Thus, we propose a fully automatic LLM-based framework to construct attack graphs named: AttacKG+. Our framework consists of four consecutive modules: rewriter, parser, identifier, and summarizer, each of which is implemented by instruction prompting and in-context learning empowered by LLMs. Furthermore, we upgrade the existing attack knowledge schema and propose a comprehensive version. We represent a cyber attack as a temporally unfolding event, each temporal step of which encapsulates three layers of representation, including behavior graph, MITRE TTP labels, and state summary. Extensive evaluation demonstrates that: (1) our formulation seamlessly satisfies the information needs in threat event analysis, (2) our construction framework is effective in faithfully and accurately extracting the information defined by AttacKG+. and (3) our attack graph directly benefits downstream security practices such as attack reconstruction. All the code and datasets will be released upon acceptance.

FluencyBank Timestamped: An Updated Data Set for Disfluency Detection and Automatic Intended Speech Recognition.

This work introduces updated transcripts, disfluency annotations, and word timings for FluencyBank, which we refer to as FluencyBank Timestamped. This data set will enable the thorough analysis of how speech processing models (such as speech recognition and disfluency detection models) perform when evaluated with typical speech versus speech from people who stutter (PWS). We update the FluencyBank data set, which includes audio recordings from adults who stutter, to explore the robustness of speech processing models. Our update (semi-automated with manual review) includes new transcripts with timestamps and disfluency labels corresponding to each token in the transcript. Our disfluency labels capture typical disfluencies (filled pauses, repetitions, revisions, and partial words), and we explore how speech model performance compares for Switchboard (typical speech) and FluencyBank Timestamped. We present benchmarks for three speech tasks: intended speech recognition, text-based disfluency detection, and audio-based disfluency detection. For the first task, we evaluate how well Whisper performs for intended speech recognition (i.e., transcribing speech without disfluencies). For the next tasks, we evaluate how well a Bidirectional Embedding Representations from Transformers (BERT) text-based model and a Whisper audio-based model perform for disfluency detection. We select these models, BERT and Whisper, as they have shown high accuracies on a broad range of tasks in their language and audio domains, respectively. For the transcription task, we calculate an intended speech word error rate (isWER) between the model's output and the speaker's intended speech (i.e., speech without disfluencies). We find isWER is comparable between Switchboard and FluencyBank Timestamped, but that Whisper transcribes filled pauses and partial words at higher rates in the latter data set. Within FluencyBank Timestamped, isWER increases with stuttering severity. For the disfluency detection tasks, we find the models detect filled pauses, revisions, and partial words relatively well in FluencyBank Timestamped, but performance drops substantially for repetitions because the models are unable to generalize to the different types of repetitions (e.g., multiple repetitions and sound repetitions) from PWS. We hope that FluencyBank Timestamped will allow researchers to explore closing performance gaps between typical speech and speech from PWS. Our analysis shows that there are gaps in speech recognition and disfluency detection performance between typical speech and speech from PWS. We hope that FluencyBank Timestamped will contribute to more advancements in training robust speech processing models.

Zeolite Stone Buddha Sculptures (pulsŏkche) in the Kyŏngsang Province during the Late Chosŏn Dynasty: Production and Transportation

The rising popularity of zeolite stone Buddhist sculptures (pulsŏkche) during the late Chosŏn Dynasty meant that temples and monk-artisans encountered new challenges. The material properties of stone, difficult to transport and limited by regional availability, posed logistical difficulties. Unlike traditional materials like wood or clay, which artisans could work with directly at the temple, stone Buddha statues often required production at a separate location from their enshrinement site. Dedicated workshops, referred to as pulsangso 佛像所 (Buddha statue production site) and chosangch'ŏ 造像處 (statue manufacturing place), were established near stone quarries for the purpose of producing these statues. This indicates that artisans likely completed the sculpting process in these workshops, particularly in ones near Kyŏngju, before transporting the sculptures to their final destinations for eye-opening ceremonies and installation in the main hall. This separation between production and enshrinement reshaped the roles and task delegation among monk-artisan groups. The responsibilities of the chŭngmyŏng (證明, the monk overseeing Buddhist projects and the ritual of depositing votive objects) became increasingly limited in stone Buddha production. On the other hand, monks involved in the image-making process took on a broader range of tasks, from procuring the raw material to carving, to depositing offerings inside the statues. Unlike those who primarily worked with wood and managed the entire production of Buddha images, pedestals, and votive plaques, the monk-artisans working with zeolite stone focused exclusively on the creation of the stone images.

Unsupervised, piecewise linear decoding enables an accurate prediction of muscle activity in a multi-task brain computer interface.

Creating an intracortical brain-computer interface (iBCI) capable of seamless transitions between tasks and contexts would greatly enhance user experience. However, the nonlinearity in neural activity presents challenges to computing a global iBCI decoder. We aimed to develop a method that differs from a globally optimized decoder to address this issue. We devised an unsupervised approach that relies on the structure of a low-dimensional neural manifold to implement a piecewise linear decoder. We created a distinctive dataset in which monkeys performed a diverse set of tasks, some trained, others innate, while we recorded neural signals from the motor cortex (M1) and electromyographs (EMGs) from upper limb muscles. We used both linear and nonlinear dimensionality reduction techniques to discover neural manifolds and applied unsupervised algorithms to identify clusters within those spaces. Finally, we fit a linear decoder of EMG for each cluster. A specific decoder was activated corresponding to the cluster each new neural data point belonged to. We found clusters in the neural manifolds corresponding with the different tasks or task sub-phases. The performance of piecewise decoding improved as the number of clusters increased and plateaued gradually. With only two clusters it already outperformed a global linear decoder, and unexpectedly, it outperformed even a global recurrent neural network (RNN) decoder with 10-12 clusters. This study introduced a computationally lightweight solution for creating iBCI decoders that can function effectively across a broad range of tasks. EMG decoding is particularly challenging, as muscle activity is used, under varying contexts, to control interaction forces and limb stiffness, as well as motion. The results suggest that a piecewise linear decoder can provide a good approximation to the nonlinearity between neural activity and motor outputs, a result of our increased understanding of the structure of neural manifolds in motor cortex.

Sustaining attention in visuomotor timing is associated with location-based binding

Maintaining focus of attention over prolonged periods can be challenging, especially when the target stimulus is absent from the temporal sequence. Prior research has shown that a temporal attentional cue filling in the temporal blank can improve sustained attention: in a sustained visual attention task requiring synchronizing finger tapping with a temporally regular sequence composed of brief flash disks interleaved with blank periods, task performance was improved when a continuous fixation point that served as a temporal attentional cue was presented superimposed on the disk stimulus. To test the hypothesis that binding the temporal attentional cue with the target temporal sequence by spatial overlapping is crucial for enhancing sustained attention, the present study conducted a series of three experiments that deconstructed the bound connection between the cue and the sequence stimulus. In Experiment 1, the cue was placed above or below a flash disk. In Experiment 2, the cue was between two vertically arranged flash disks. In Experiment 3, the cue was in a flash ring. No significant effect of sustained attention improvement was found in any of the three experiments. Experiment 4 further replicated these null results and the previously observed effect of sustained attention improvement when the temporal cue was superimposed on the sequence stimulus. Our finding demonstrates that binding by spatial overlapping during the temporal blank when the sequence stimulus is absent is critical for enhancing sustained attention, which should be beneficial for improving performance across a broader range of tasks that require prolonged maintenance of attention.

EviPrompt: A Training-Free Evidential Prompt Generation Method for Adapting Segment Anything Model in Medical Images.

Medical image segmentation is a critical task in clinical applications. Recently, the Segment Anything Model (SAM) has demonstrated potential for natural image segmentation. However, the requirement for expert labour to provide prompts, and the domain gap between natural and medical images pose significant obstacles in adapting SAM to medical images. To overcome these challenges, this paper introduces a novel prompt generation method named EviPrompt. The proposed method requires only a single reference image-annotation pair, making it a training-free solution that significantly reduces the need for extensive labelling and computational resources. First, prompts are automatically generated based on the similarity between features of the reference and target images, and evidential learning is introduced to improve reliability. Then, to mitigate the impact of the domain gap, committee voting and inference-guided in-context learning are employed, generating prompts primarily based on human prior knowledge and reducing reliance on extracted semantic information. EviPrompt represents an efficient and robust approach to medical image segmentation. We evaluate it across a broad range of tasks and modalities, confirming its efficacy. The source code is available at https://github.com/SPIresearch/EviPrompt.

The impact of an active and passive industrial back exoskeleton on functional performance

Due to differences in actuation and design, active and passive industrial back exoskeletons could influence functional performance, i.e., work performance, perceived task difficulty, and discomfort, differently. Therefore, this study investigated and compared the impact of the active CrayX (7 kg) and passive Paexo Back (4.5 kg) on functional performance. Eighteen participants performed twelve work-related tasks with both types of exoskeletons and without (NoExo). The CrayX hindered work performance up to 22% in multiple tasks, compared to the Paexo Back and NoExo, while work performance between NoExo and the Paexo Back condition was more comparable, except for stair climbing (13% hindrance). Perceived task difficulty and discomfort seldomly varied between both exoskeletons. Although the CrayX shows promise to benefit workers, limitations in hindrance and comfort should first be addressed. The Paexo Back has demonstrated an advantage in certain static tasks. However, increasing its potential across a broader range of tasks seems warranted. Practitioner Summary: Differences between industrial back exoskeletons with regard to functional performance, i.e. work performance, discomfort and perceived task difficulty, were investigated by evaluating the active CrayX and passive Paexo Back back exoskeletons. The CrayX significantly hindered functional performance, while the Paexo Back seldomly affected functional performance. Abbreviations: WMSD: Work-related musculoskeletal disorder; NoExo: No Exoskeleton; GD: General discomfort; PTD: Perceived task difficulty; BMI: Body Mass Index

What does a Peer Support Worker do in a Forensic Mental Health Clinic for Addicted Offenders?

IntroductionPeer support work can be an effective way to support patients and their participation in psychiatric treatment. Unlike in general psychiatry there is less experience with peer support work in forensic mental health inpatient settings. Characteristics different from general psychiatry, e.g., regarding patient structure and background, might lead to different tasks of peer support workers and subjects of conversation.ObjectivesWe aim to present an overview of tasks and conversation topics of a peer support worker in a forensic mental health setting for addicted offenders during an 12 month period. We address tasks on regular and irregular basis and the most frequent conversation topics.MethodsWe used the anonymized work documentation about weekly working activities and conversation notes of the peer support worker to extract information. Extracted data were thematically analyzed and clustered into themes for tasks and conversation topics.ResultsResults reveal several recurring and routine tasks, like joining ward rounds and changes of shift or leading a recovery patient group along with one-to-one conversations with patients according to their request. These topics were expanded by irregular tasks like group discussions for special occasions, e.g., after incidents. During one-on-one conversations, patients addressed topics, e.g., about their substance use history, thoughts and issues about their therapy or ways to achieve their goals in the future.ConclusionsPeer support is a well-accepted offer that can contain various different tasks in groups and in one-on-one settings as well. Although the peer support worker is seated on one ward, there are many requests from other wards, too. There also exists a broad range of conversation topics, some might be also present in general psychiatric wards and others that might be more unique to forensic settings. The broad range of tasks and acceptance of peer support make it necessary to provide corresponding resources like peer support staff and payment.Disclosure of InterestNone Declared

Signatures of contextual interference in implicit sensorimotor adaptation.

Contextual interference refers to the phenomenon whereby a blocked practice schedule results in faster acquisition but poorer retention of new motor skills compared to a random practice schedule. While contextual interference has been observed under a broad range of tasks, it remains unclear if this effect generalizes to the implicit and automatic recalibration of an overlearned motor skill. To address this question, we compared blocked and random practice schedules in a visuomotor rotation task that isolates implicit adaptation. In experiment 1, we found robust signatures of contextual interference in implicit adaptation: compared to participants tested under a blocked training schedule, participants tested under a random training schedule exhibited a reduced rate of learning during the training phase but better retention during a subsequent no-feedback assessment phase. In experiment 2, we again observed an advantage in retention following random practice and showed that this result was not due to a change in context between the training and assessment phases (e.g. a blocked training schedule followed by a random assessment schedule). Taken together, these results indicate that contextual interference is not limited to the acquisition of new motor skills but also applies to the implicit adaptation of established motor skills.

Influence of working conditions and salary on agency work for intermediate and intensive care units : Part2: Partial results of a nationwide survey.

Agency work in nursing is used as aform of labor to counter vacant staff positions in hospitals. Both hospital owners and nurses view this critically for different reasons. The aim of this study was to assess what personal net income nurses in German intensive care units and intermediate care units consider "fair and sufficient" for their work (addressed in Part1 of the survey) and what influence-aside from the salary-the working conditions have on the willingness to change to temporary work or back to apermanent position. From September to October 2020, an anonymous online survey was conducted among nurses of intermediate care units, intensive care units, and special care units in German-speaking countries. Descriptive statistics were used for the analysis. Of 1203 participants, 86% (n = 1036) could be evaluated. None of the job satisfaction factors queried received four or five stars (maximum five stars) from those participating in the survey. The most unsatisfied group proved to be regularly employed nurses with an additional part-time job. Key job satisfaction factors differed markedly between the groups, with regular employees favoring consistency and stability. Agency workers prefer gaining experience in abroader range of tasks. Unreliable duty rosters and poor nurse to patient ratios were common points of criticism. For job satisfaction, making nurses feel appreciated and respected is essential. This includes aguaranteed nurse to patient ratio and reliable duty rosters that also include tasks outside direct patient care. In order for nurses to leave agency work, it is necessary to take into account the differences in interests in terms of the focus of activity.

Multi-spectral multi-image super-resolution of Sentinel-2 with radiometric consistency losses and its effect on building delineation

High resolution remote sensing imagery is used in a broad range of tasks, including detection and classification of objects. High-resolution imagery is however expensive to obtain, while lower resolution imagery is often freely available and can be used for a range of social good applications. To that end, we curate a multi-spectral multi-image dataset for super-resolution of satellite images. We use PlanetScope imagery from the SpaceNet-7 challenge as the high resolution reference and multiple Sentinel-2 revisits of the same location as the low-resolution imagery. We present the first results of applying multi-image super-resolution (MISR) to multi-spectral remote sensing imagery. We, additionally, introduce a radiometric-consistency module into the MISR model to preserve the high radiometric resolution and quality of the Sentinel-2 sensor. We show that MISR is superior to single-image super-resolution (SISR) and other baselines on a range of image fidelity metrics. Furthermore, we present the first assessment of the utility of multi-image super-resolution on a semantic and instance segmentation – common remote sensing tasks – showing that utilizing multiple images results in better performance in these downstream tasks, but MISR pre-processing is non-essential.

Gender and creative ability: Mean differences and performance variability

Studies examining gender differences in creative abilities remain inconsistent and call for further research using diverse measures. The current study examined gender differences in mean performance, as well as differences in variability on different performance criteria for male and female participants. The study includes a sample of children and early-to-mid adolescents (N = 678; 1st–10th grade in primary and secondary schools) and uses tasks in verbal (storytelling) and figural domains (incomplete figures) scored for six performance criteria (creativity, novelty/originality, likability, elaboration, humor, and emotion use). Similar to previous research, when mean differences in performance emerged, they favored girls; however, these differences tended to be small and not consistent across all performance criteria. The greater male variability hypothesis in performance was not supported for any of the performance measures. These results point to the limits of the greater male variability hypothesis and the need to examine gender differences in creative abilities across a broad range of tasks and domains and investigate developmental factors that could be related to the emergence of gender differences.

Robotics and artificial intelligence in healthcare during COVID-19 pandemic

This paper will present the use of robotics and artificial intelligence in healthcare during COVID-19 pandemic. The types of robots to be discussed are autonomous mobile platforms that can assist with a broad range of tasks, including remote monitoring, on-site rescue and logistics. A case study is presented to illustrate how an artificial intelligence network application enabled healthcare providers to plan for the staffing levels of specific care areas while taking into account other constraints such as security and connectivity issues during a pandemic event. Robotics have become ubiquitous in human society in recent years due to their ability to perform repetitive tasks with minimal supervision, thus freeing up humans for more important activities or providing them relief from monotonous work that is physically demanding or exhausting.

Predicting brain activation maps for arbitrary tasks with cognitive encoding models

A deep understanding of the neural architecture of mental function should enable the accurate prediction of a specific pattern of brain activity for any psychological task, based only on the cognitive functions known to be engaged by that task. Encoding models (EMs), which predict neural responses from known features (e.g., stimulus properties), have succeeded in circumscribed domains (e.g., visual neuroscience), but implementing domain-general EMs that predict brain-wide activity for arbitrary tasks has been limited mainly by availability of datasets that 1) sufficiently span a large space of psychological functions, and 2) are sufficiently annotated with such functions to allow robust EM specification. We examine the use of EMs based on a formal specification of psychological function, to predict cortical activation patterns across a broad range of tasks. We utilized the Multi-Domain Task Battery, a dataset in which 24 subjects completed 32 ten-minute fMRI scans, switching tasks every 35s and engaging in 44 total conditions of diverse psychological manipulations. Conditions were annotated by a group of experts using the Cognitive Atlas ontology to identify putatively engaged functions, and region-wise cognitive EMs (CEMs) were fit, for individual subjects, on neocortical responses. We found that CEMs predicted cortical activation maps of held-out tasks with high accuracy, outperforming a permutation-based null model while approaching the noise ceiling of the data, without being driven solely by either cognitive or perceptual-motor features. Hierarchical clustering on the similarity structure of CEM generalization errors revealed relationships amongst psychological functions. Spatial distributions of feature importances systematically overlapped with large-scale resting-state functional networks (RSNs), supporting the hypothesis of functional specialization within RSNs while grounding their function in an interpretable data-driven manner. Our implementation and validation of CEMs provides a proof of principle for the utility of formal ontologies in cognitive neuroscience and motivates the use of CEMs in the further testing of cognitive theories.

Fully body visual self-modeling of robot morphologies.

Internal computational models of physical bodies are fundamental to the ability of robots and animals alike to plan and control their actions. These "self-models" allow robots to consider outcomes of multiple possible future actions without trying them out in physical reality. Recent progress in fully data-driven self-modeling has enabled machines to learn their own forward kinematics directly from task-agnostic interaction data. However, forward kinematic models can only predict limited aspects of the morphology, such as the position of end effectors or velocity of joints and masses. A key challenge is to model the entire morphology and kinematics without prior knowledge of what aspects of the morphology will be relevant to future tasks. Here, we propose that instead of directly modeling forward kinematics, a more useful form of self-modeling is one that could answer space occupancy queries, conditioned on the robot's state. Such query-driven self-models are continuous in the spatial domain, memory efficient, fully differentiable, and kinematic aware and can be used across a broader range of tasks. In physical experiments, we demonstrate how a visual self-model is accurate to about 1% of the workspace, enabling the robot to perform various motion planning and control tasks. Visual self-modeling can also allow the robot to detect, localize, and recover from real-world damage, leading to improved machine resiliency.

What Is Missing from Contemporary AI? The World

In the past three years, we have witnessed the emergence of a new class of artificial intelligence systems–—so-called foundation models, which are characterised by very large machine learning models (with tens or hundreds of billions of parameters) trained using extremely large and broad data sets. Foundation models, it is argued, have competence in a broad range of tasks, which can be specialised for specific applications. Large language models, of which GPT-3 is perhaps the best known, are the most prominent example of current foundation models. While foundation models have demonstrated impressive capabilities in certain tasks—natural language generation being the most obvious example—I argue that because they are inherently disembodied, and they are limited with respect to what they have learned and what they can do. Foundation models are likely to be very useful in many applications: but they are not the end of the road in artificial intelligence.

The Task Matters: A Scoping Review on Reading Comprehension Abilities in ADHD

Objective: A broad range of tasks have been used to classify individuals with ADHD with reading comprehension difficulties. However, the inconsistency in the literature warrants a scoping review of current knowledge about the relationship between ADHD diagnosis and reading comprehension ability. Method: A comprehensive search strategy was performed to identify relevant articles on the topic. Thirty-four articles met inclusion criteria for the current review. Results: The evidence as a whole suggests reading comprehension is impaired in ADHD. The most prominent effect was found in studies where participants retell or pick out central ideas in stories. On these tasks, participants with ADHD performed consistently worse than typically developing controls. However, some studies found that performance in ADHD improved when reading comprehension task demands were low. Conclusion: Results suggest that performance in ADHD depends on the way reading comprehension is measured and further guide future work clarifying why there are such discrepant findings across studies.

Changes of Neuronal Plasticity in Postoperative Cognitive Dysfunction (POCD)

AbstractBackgroundPostoperative cognitive dysfunction (POCD) is defined as a prolonged cognitive function impairment that occurs within weeks to months of a surgical procedure. Clinically this is a well‐known risk of the surgical experience in the elderly, having been described as a consequence of anesthesia. The medial prefrontal cortex (mPFC) plays a critical role in remote, recent, and short‐term memories over a broad range of tasks. The effect of POCD on neuronal plasticity in mPFC remains controversial. The present study investigates the impact of laparotomy on cognition and synaptic functions.MethodMice were induced with 5% sevoflurane and maintained with 3 ‐ 4% sevoflurane continuously for 30 mins. Under sevoflurane anesthesia, a 2.5 cm longitudinal midline incision was made. Approximately 10 cm of intestine was exteriorized and vigorously rubbed for 2 min and stay outside the abdominal cavity for 1 min. Novel Object Recognition (NOR) test was performed 1hour and 24 hours for the short‐term and long‐term memory seven days after the surgery. Open field test (OFT) and The Elevated Plus Maze test, was performed to exclude motor dysfunction and anxiety. The miniature excitatory postsynaptic currents (mEPSC) and miniature inhibitory postsynaptic currents (mIPSC) were recorded on the 7 or 14 days after surgery.ResultIn the NOR test, mice with undergone laparotomy surgery had worse cognitive performance than the control group. There was no significance between the laparotomy group and the control group in the OFT test and Elevated Plus Maze test. There was no significance in the amplitude of the mEPSC between laparotomy and control group. However, the frequency increased significantly in the laparotomy group than the control group, which can last for 14 days. There was no significance of mIPSC in the amplitude and frequency between the laparotomy group and the control group.ConclusionThese findings demonstrated that laparotomy surgery under sevoflurane leads to learning and memory impairment in adolescent mice without altering motor dysfunction and anxiety. The increase in the presynapse release ability indicated the glutamatergic neuron abnormalities without altering the GABAergic neuron’s function. Our result suggested that postoperative cognitive dysfunctions may be related to surgical trauma and glutamate neurons abnormalities.

Perturbation-based methods for explaining deep neural networks: A survey

• Overview of the black box problem in DNNs and XAI research. • Survey of perturbation-based attribution methods for different input data types. • Future directions for research on perturbation paradigm. Deep neural networks (DNNs) have achieved state-of-the-art results in a broad range of tasks, in particular the ones dealing with the perceptual data. However, full-scale application of DNNs in safety-critical areas is hindered by their black box-like nature, which makes their inner workings nontransparent. As a response to the black box problem, the field of explainable artificial intelligence (XAI) has recently emerged and is currently rapidly growing. The present survey is concerned with perturbation-based XAI methods, which allow to explore DNN models by perturbing their input and observing changes in the output. We present an overview of the most recent research focusing on the differences and similarities in the applications of perturbation-based methods to different data types, from extensively studied perturbations of images to the just emerging research on perturbations of video, natural language, software code, and reinforcement learning entities.

Deep neural networks using a single neuron: folded-in-time architecture using feedback-modulated delay loops

Deep neural networks are among the most widely applied machine learning tools showing outstanding performance in a broad range of tasks. We present a method for folding a deep neural network of arbitrary size into a single neuron with multiple time-delayed feedback loops. This single-neuron deep neural network comprises only a single nonlinearity and appropriately adjusted modulations of the feedback signals. The network states emerge in time as a temporal unfolding of the neuron’s dynamics. By adjusting the feedback-modulation within the loops, we adapt the network’s connection weights. These connection weights are determined via a back-propagation algorithm, where both the delay-induced and local network connections must be taken into account. Our approach can fully represent standard Deep Neural Networks (DNN), encompasses sparse DNNs, and extends the DNN concept toward dynamical systems implementations. The new method, which we call Folded-in-time DNN (Fit-DNN), exhibits promising performance in a set of benchmark tasks.