Search
Results
-
Injecting Descriptive Meta-Information into Pre-Trained Language Models with Hypernetworks
(2021)Proceedings of Interspeech 2021There is a growing trend to deploy deep neural networks at the edge for high-accuracy, real-time data mining and user interaction. Applications such as speech recognition and language understanding often apply a deep neural network to encode an input sequence and then use a decoder to generate the output sequence. A promising technique to accelerate these applications on resource-constrained devices is network pruning, which compresses ...Conference Paper -
Automatic Energy-Hotspot Detection and Elimination in Real-Time Deeply Embedded Systems
(2021)2021 IEEE Real-Time Systems Symposium (RTSS)Today’s deeply embedded systems, with real-time interactions to the environment, are largely battery-operated, and peripheral modules like LTE, WiFi, and GPS are among the most energy-hungry components of them. These components are often under the direct control of an embedded software. Some pieces of the software program are called energy hotspots if they can be transformed towards better system energy consumption while leaving it ...Conference Paper -
TIP-Air: Tracking Pollution Transfer for Accurate Air Quality Prediction
(2021)Adjunct Proceedings of the 2021 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2021 ACM International Symposium on Wearable Computers (UbiComp-ISWC '21 Adjunct)Air quality is of vital importance to human health. Accurately predicting air quality, especially its sudden changes, is highly valuable for citizens and governments to make personal and local decisions, design intelligent policies and control pollution at minimal cost. However, none of the existing methods achieves sufficient prediction accuracy for time intervals of sudden pollution change due to inability of existing models to take ...Conference Paper -
Pruning-Aware Merging for Efficient Multitask Inference
(2021)Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery & Data Mining (KDD '21)Many mobile applications demand selective execution of multiple correlated deep learning inference tasks on resource-constrained platforms. Given a set of deep neural networks, each pre-trained for a single task, it is desired that executing arbitrary combinations of tasks yields minimal computation cost. Pruning each network separately yields suboptimal computation cost due to task relatedness. A promising remedy is to merge the networks ...Conference Paper -
HiMap: Fast and Scalable High-Quality Mapping on CGRA via Hierarchical Abstraction
(2021)Proceedings of the 2021 Design, Automation & Test in Europe (DATE 2021)Coarse-Grained Reconfigurable Array (CGRA) has emerged as a promising hardware accelerator due to the excellent balance among reconfigurability, performance, and energy efficiency. The CGRA performance strongly depends on a high-quality compiler to map the application kernels on the architecture. Unfortunately, the state-of-the-art compilers fall short in generating high quality mapping within an acceptable compilation time, especially ...Conference Paper -
Optimal Power Management for Energy Harvesting Systems with A Backup Power Source
(2021)2021 10th Mediterranean Conference on Embedded Computing (MECO)Energy harvesting has been extensively used to allow for long-term and unattended operation of nodes in large-scale distributed systems. However, the smaller the relative rechargeable energy storage of the harvesting system, the higher is the sensitivity of its operation to short-term non-deterministic changes of the harvested power or the current power demand. A reliable and predictable node operation can be achieved with an additional ...Conference Paper -
Using system context information to complement weakly labeled data
(2021)arXiv ~ Proceedings of the First Workshop on Weakly Supervised Learning (WeaSuL), May 7, 2021, co-located with ICLR (Online)Conference Paper -
Pruning Meta-Trained Networks for On-Device Adaptation
(2021)Proceedings of the 30th ACM International Conference on Information & Knowledge ManagementAdapting neural networks to unseen tasks with few training samples on resource-constrained devices benefits various Internet-of-Things applications. Such neural networks should learn the new tasks in few shots and be compact in size. Meta-learning enables few-shot learning, yet the meta-trained networks can be over-parameterised. However, naive combination of standard compression techniques like network pruning with meta-learning jeopardises ...Conference Paper -
STeC: Exploiting Spatial and Temporal Correlation for Event-based Communication in WSNs
(2021)SenSys '21: Proceedings of the 19th ACM Conference on Embedded Networked Sensor SystemsLow-power wireless sensor networks have demonstrated their potential for the detection of rare events such as rockfalls and wildfires, where rapid reporting as well as long-term energy-efficient operation is vital. However, current systems require periodic synchronization to maintain network coordination, heavily rely on node placement or use costly long-range links to infrastructure. We present STeC, a novel wireless communication design ...Conference Paper -
Joint Energy Management for Distributed Energy Harvesting Systems
(2021)SenSys '21: Proceedings of the 19th ACM Conference on Embedded Networked Sensor SystemsConference Paper