Results 241 to 250 of about 10,598,544 (298)

MMLU-Pro: A More Robust and Challenging Multi-Task Language Understanding Benchmark

Neural Information Processing Systems
In the age of large-scale language models, benchmarks like the Massive Multitask Language Understanding (MMLU) have been pivotal in pushing the boundaries of what AI can achieve in language comprehension and reasoning across diverse domains.
Yubo Wang, Xueguang Ma, Ge Zhang, Yuansheng Ni, Abhranil Chandra, Shiguang Guo, Weiming Ren, Aaran Arulraj, Xuan He, Ziyan Jiang, Tianle Li, Max W.F. Ku, Kai Wang, Alex Zhuang, Rongqi "Richard" Fan, Xiang Yue, Wenhu Chen   +16 more
semanticscholar   +1 more source

Tasks for and tasks in human–computer interaction

Interacting with Computers, 2006
It is argued that the engineering discipline of human-computer interaction (HCI) has developed over the last couple of decades in an ad hoc manner, driven by the need to solve real problems, rather than being informed by high level, general theories of the discipline. A retrospective role for general HCI theories is suggested. A start at such a general
Dan Diaper, Colston Sanger
openaire   +1 more source

Task complexity and task type

TASK. Journal on Task-Based Language Teaching and Learning, 2023
Abstract In Task-Based Language Teaching (TBLT), there is clearly a need for a balanced and flexible view of L1 use. Considering this, a matter of concern in TBLT is that learners might overuse their first language (L1) while performing complex tasks.
Mahnaz Entezari, Mojtaba Tadayonifar
openaire   +1 more source

TOFU: A Task of Fictitious Unlearning for LLMs

arXiv.org
Large language models trained on massive corpora of data from the web can memorize and reproduce sensitive or private data raising both legal and ethical concerns.
Pratyush Maini, Zhili Feng, Avi Schwarzschild, Zachary Chase Lipton, J. Kolter   +4 more
semanticscholar   +1 more source

Multi-task Learning with Task Relations

2011 IEEE 11th International Conference on Data Mining, 2011
Multi-task and relational learning with Gaussian processes are two active but also orthogonal areas of research. So far, there has been few attempt at exploring relational information within multi-task Gaussian processes. While existing relational Gaussian process methods have focused on relations among entities and in turn could be employed within an ...
Zhao Xu 0001, Kristian Kersting
openaire   +1 more source

Task preparation and task repetition: Two-component model of task switching.

Journal of Experimental Psychology: General, 2001
The switch cost (the disadvantage of performing a new task vs. a repeated task) has been attributed to lack of preparation for the switched task or priming of the repeated task. These sources were examined by manipulating foreknowledge of task transition (repeat or switch), response-to-stimulus interval (RSI), and practice level.
M H, Sohn, J R, Anderson
openaire   +2 more sources

Task Shifting in Dual-Task Settings

Perceptual and Motor Skills, 2002
When a participant is asked to perform two tasks in alternation, their mean reaction times were slower than when they performed the same tasks repeatedly. This “shift cost” has been hypothesized to reflect the time course of a single central executive that exerts control over thought and actions in task shifting.
openaire   +2 more sources

End-To-End Multi-Task Learning With Attention

Computer Vision and Pattern Recognition, 2018
We propose a novel multi-task learning architecture, which allows learning of task-specific feature-level attention. Our design, the Multi-Task Attention Network (MTAN), consists of a single shared network containing a global feature pool, together with ...
Shikun Liu, Edward Johns, A. Davison
semanticscholar   +1 more source

Problems in modelling tasks and task views

Conference Sponsored by ACM SIGOIS and IEEECS TC-OA on Office information systems -, 1988
We present the rudiments of a new theory of tasks and task views. The motivating idea is essentially as follows: in an environment in which several organizational agents cooperate to accomplish a common task, each of the agents need only know its own part of the task — this is the agent's task view.
openaire   +1 more source

Subjective Estimation of Task Time and Task Difficulty of Simple Movement Tasks

Journal of Motor Behavior, 2016
It has been demonstrated in previous work that the same neural structures are used for both imagined and real movements. To provide a strong test of the similarity of imagined and actual movement times, 4 simple movement tasks were used to determine the relationship between estimated task time and actual movement time.
Alan H S, Chan, Errol R, Hoffmann
openaire   +2 more sources