Environmental impact assessment (EIA), a decision-making process for project appraisal and sustainability adopted globally as an administrative process to identify, predict, evaluate, and monitor projects from their f...Environmental impact assessment (EIA), a decision-making process for project appraisal and sustainability adopted globally as an administrative process to identify, predict, evaluate, and monitor projects from their feasibility, preconstruction, construction, and operation stages to mitigate the adverse impacts and enhance the beneficial impacts for the protection of the affected environment, The study objective is to explore global EIA systems and processes and find shortcomings and implications for making the best instrument or tool to protect the natural environment from man-made activities over the project cycle. For this, the relevant literature on the EIA system and process was reviewed and evaluated through the application of quantitative and qualitative approaches, including the assessment of legal instruments and the adoption of EIA methodologies in developing countries. EIA, initiated in the 1970s in the US and Australia, expanded to developing countries and was amended in East and Southeast Asia from the 1970s to the 2000s. The evaluation assessed that the South Asian countries follow UNEP and IAIA guidelines, utilizing national laws and expert consultations, with screening criteria and administrative processes based on established environmental legislation. Ad hoc, checklist, matrix, network, overlay, cost-benefit analysis, and predictive or simulation in EIA practice are used to assess the environmental impacts of development activities. Failure to recommend major projects undermines public trust and prevents mitigation measures from being implemented. Most developing countries have followed EIA to fulfill the legal requirement with shadow-off monitoring and follow-up rather than to upset, reduce, or compensate for the project impacts as per size, location, and severity of the project area. The research and guidelines outlined in the IAIA principles and process have synthesized the best EIA practices worldwide. Public participation, impact coverage, scientific mitigation, transparent evidence-based approaches, monitoring, follow-up, legitimate approaches, and future appraisal opportunities are major concerns to be included in best EIA practice.展开更多
深度学习在自然语言处理方面取得了巨大进展,以深度神经网络为代表的模型开始在法律智能判决上被广泛使用。基于Transformer的双向编码器表征法(Bidirectional Encoder Representations from Transformers,BERT)模型能够挖掘法律描述文...深度学习在自然语言处理方面取得了巨大进展,以深度神经网络为代表的模型开始在法律智能判决上被广泛使用。基于Transformer的双向编码器表征法(Bidirectional Encoder Representations from Transformers,BERT)模型能够挖掘法律描述文本中双向上下文信息,利用BERT中自注意力机制完成了罪名预测、法律条款推荐、刑期预测多个司法智能审判任务。为了在长文本案情描述文本上获得更好的效果,进一步解决BERT模型输入文本的长度限制,对于过长的输入文本进行关键信息提取。在文本提取的过程中,充分利用前期训练的基于BERT智能审判模型,对于案情描述中句子的重要性进行评估,提取关键句子减少判断模型的输入长度。将精简后的案情描述文本再送入BERT模型进行司法智能审判学习。相比于直接输入原始案情描述文本的方法,基于文本提取处理后的法律描述在智能审判任务中能够取得更好的效果。展开更多
文摘Environmental impact assessment (EIA), a decision-making process for project appraisal and sustainability adopted globally as an administrative process to identify, predict, evaluate, and monitor projects from their feasibility, preconstruction, construction, and operation stages to mitigate the adverse impacts and enhance the beneficial impacts for the protection of the affected environment, The study objective is to explore global EIA systems and processes and find shortcomings and implications for making the best instrument or tool to protect the natural environment from man-made activities over the project cycle. For this, the relevant literature on the EIA system and process was reviewed and evaluated through the application of quantitative and qualitative approaches, including the assessment of legal instruments and the adoption of EIA methodologies in developing countries. EIA, initiated in the 1970s in the US and Australia, expanded to developing countries and was amended in East and Southeast Asia from the 1970s to the 2000s. The evaluation assessed that the South Asian countries follow UNEP and IAIA guidelines, utilizing national laws and expert consultations, with screening criteria and administrative processes based on established environmental legislation. Ad hoc, checklist, matrix, network, overlay, cost-benefit analysis, and predictive or simulation in EIA practice are used to assess the environmental impacts of development activities. Failure to recommend major projects undermines public trust and prevents mitigation measures from being implemented. Most developing countries have followed EIA to fulfill the legal requirement with shadow-off monitoring and follow-up rather than to upset, reduce, or compensate for the project impacts as per size, location, and severity of the project area. The research and guidelines outlined in the IAIA principles and process have synthesized the best EIA practices worldwide. Public participation, impact coverage, scientific mitigation, transparent evidence-based approaches, monitoring, follow-up, legitimate approaches, and future appraisal opportunities are major concerns to be included in best EIA practice.
文摘深度学习在自然语言处理方面取得了巨大进展,以深度神经网络为代表的模型开始在法律智能判决上被广泛使用。基于Transformer的双向编码器表征法(Bidirectional Encoder Representations from Transformers,BERT)模型能够挖掘法律描述文本中双向上下文信息,利用BERT中自注意力机制完成了罪名预测、法律条款推荐、刑期预测多个司法智能审判任务。为了在长文本案情描述文本上获得更好的效果,进一步解决BERT模型输入文本的长度限制,对于过长的输入文本进行关键信息提取。在文本提取的过程中,充分利用前期训练的基于BERT智能审判模型,对于案情描述中句子的重要性进行评估,提取关键句子减少判断模型的输入长度。将精简后的案情描述文本再送入BERT模型进行司法智能审判学习。相比于直接输入原始案情描述文本的方法,基于文本提取处理后的法律描述在智能审判任务中能够取得更好的效果。
