This study explores the use of chaos engineering in evaluating the performance and resilience of serverless applications, which are built as complex distributed systems subject to different types of failures and error...
详细信息
Data-to-text (D2T) generation, the task of converting structured data into natural language, has extensive real-world applications. While supervised models have achieved promising results, they rely heavily on costly ...
详细信息
ISBN:
(纸本)9798350359329;9798350359312
Data-to-text (D2T) generation, the task of converting structured data into natural language, has extensive real-world applications. While supervised models have achieved promising results, they rely heavily on costly labeled training data. This paper investigates unsupervised D2T generation by leveraging the impressive general abilities of large language models (LLMs). We propose a framework for LLMs to collaboratively learn from unlabeled data through cascaded multi-model quality control. Specifically, one LLM, acting as a writer, generates candidate texts from input data. Additional LLMs, serving as checkers, validate output quality to filter high-quality samples for training the writer LLM. By cascading generation, checking, and meta-checking, the models extract linguistic knowledge and grounding ability from abundant unlabeled data. Experiments on established benchmarks demonstrate enhanced fluency, accuracy, and coherence compared to supervised baselines. This unsupervised approach circumvents labeled data dependence, unlocking readily available LLMs for on-demand D2T generation across diverse applications.
In recent years, with the support of appropriate smart contract-based access control policies, blockchain technology has proven to be a compelling solution for providing a unified, trusted platform for resource sharin...
详细信息
ISBN:
(纸本)9798350310191
In recent years, with the support of appropriate smart contract-based access control policies, blockchain technology has proven to be a compelling solution for providing a unified, trusted platform for resource sharing. However, due to the immutability property of blockchain, it can be challenging to patch or fix bugs after the deployment of a smart contract. Therefore, it is critical to ensure that the smart contract access control policies adhere to all specifications prior to their deployment, and the presence of anomalies within policies is not desirable. Thus, this paper proposes an automated policy verification and enforcement framework for Ethereum decentralized applications, which supports an automated off-chain policy verification before its deployment to the underlying blockchain. We integrate a verification engine within the off-chain module paired with a translator to convert XACML policies into Solidity smart contracts and vice-versa. Additionally, the experiments performed on benchmark XACML policies further reinforce the pragmatism of our proposal.
The power quality of the grid is a major issue for electrical power system. Thus, power factor and harmonic profile of the current must be improved to comply with IEC standards. In this paper, a boost converter is emp...
详细信息
The squirrel cage induction machine (SCIM) based water pumping system (WPS) remains a cost-effective and robust option compared to other alternatives. However, the voltage-to-frequency-based controllers and direct-on-...
详细信息
Lighting in commercial buildings consumes a substantial amount of energy. Therefore, this paper developed particle swarm optimization (PSO) and firefly algorithm (FA) as control techniques for lighting systems to impr...
详细信息
In this study, an effective flux-weakening control for permanent magnet synchronous motor (PMSM) is proposed, which combines model predictive control (MPC) with vector control. On the one hand, in order to achieve ful...
详细信息
This paper investigates the interaction effects of permanent magnet (PM) tolerances and rotor eccentricities of 12-slot/10-pole PM machines by finite element method (FEM) and phasor analysis. It is found that the rela...
详细信息
This paper investigates the interaction effects of permanent magnet (PM) tolerances and rotor eccentricities of 12-slot/10-pole PM machines by finite element method (FEM) and phasor analysis. It is found that the relative location between diversified PMs and eccentric rotor have a significant effect on cogging torque. (a) When a weak (e.g., thin thickness) PM is close to the minimum airgap, rotor static eccentricity and PM tolerances have a strengthening effect to deteriorate cogging torque, but rotor dynamic eccentricity and PM tolerances have a weakening effect to reduce cogging torque. (b) When a strong (e.g., thick thickness) PM is close to the minimum airgap, rotor static/dynamic eccentricity and PM tolerances always have a strengthening effect to deteriorate cogging torque. The experiments verify the FEM-predicted cogging torques and the interaction effects between PM tolerances and rotor eccentricities on cogging torque in the 12-slot/10-pole PM machine.
Effective biodegradable waste management is a significant challenge in urban environments, particularly with high-volume organic waste like tender coconut shells. This research proposes an innovative IoT-based system ...
详细信息
According to the force and motion analysis principle of vertical flight, roll flight, pitch flight and yaw flight of quad rotor UAV, the mathematical model is established, the appropriate reference coordinate system i...
详细信息
暂无评论