线上讲座 | 原则AI赋能语言技能发展;计算神经语言学论坛:神经表征驱动的脑机对齐与模型增强等

读书使人进步，无暇也要读书。关注我们吧！

‍

欢迎关注我们，一站式获取海量学术资源

转载来源：香港教育大学中国语言学系、语言学之妙公众号

我们的宗旨：助研助学，促进交流

敬请星标应用语言学研习，喜欢请点赞、在看、转发⭐

线上讲座 | 计算神经语言学论坛：神经表征驱动的脑机对齐与模型增强

6月22日（周一）

（1）神经表征驱动的脑机对齐与模型增强

主办方：计算神经语言学论坛

主讲人：陈佳炫（浙江大学计算机科学与技术学院）

时间：10:00

线上观看方式：Zoom95891053788（密码213165）

讲座内容：

近年来，人工智能的发展几乎都遵循着同一个逻辑:参数更大、数据更多、模型更强，但模型虽然"知道得更多”，却未必"理解得更深”——它们可以记住海量知识，却依然会在抽象概念理解、跨场景泛化、概念关系推理等方面暴露出局限性。与此同时，以脑机接口为代表的神经技术快速发展，使得脑与机的结合愈加紧密，脑机融合及一体化逐渐成为人工智能发展的重要趋势。在这一背景下，脑机融合与增强研究受到广泛关注，成为弥补脑机行为差异、提升人工模型认知能力的关键途径之一本报告将聚焦神经表征学习、脑机表征对齐等问题，探讨神经表征驱动的脑机对齐技术及其在脑机增强中的应用，阐述其研究背景、基本概念与实现形式。

6月23日（周二）

6月24日（周三）

（1）Pixels, Prompts, and Pedagogy: AI-Enhanced Digital Multimodal Composing and Game-Based Learning as Dual Pedagogical Pathways in EFL Education

主办方：西交利物浦大学人文学院

主讲人：张丹阳（深圳大学外国语学院）

时间：10:00

线下地点：江苏省苏州市西交利物浦大学HS436

讲座内容：

The integration of generative Al into foreign language education stimulates a re-examination of pedagogical frameworks that move beyond instrumental adoption toward transformative meaning-making. This talk examines two complementary pathways-Al-enhanced Digital Multimodal Composing (DMC) and Al-enhanced Game-Based Learning (GBL)-through which AI can be applied in EFL contexts. The DMC pathway positions students as multimodal composers who design Al-generated content, drawing on Kress and van Leeuwen's (2006) visual grammar to decode how images construct experiential, interpersonal, and textual meaning. GenAI functions not as a creative substitute but as a catalyst for semiotic reflection: students craft and iterate prompts with rhetorical intent and critically evaluate GenAI output against established visual principles. Empirical findings from a mixed-methods study with 41 EFL undergraduates reveal significant gains in creation competencies, alongside meaningful improvements in analysis and evaluation, indicating that GenAI redirected cognitive effort toward credibility assessment and audience-aware design

The GBL. pathway repositions students as players and narrative architects who inhabit AI-driven story worlds. Grounded in Flow Theory and Self-Determination Theory, this strand is illustrated through a case study in which learners interrogate GenAI-powered non-player characters (NPCs) to solve a mystery. The talk details the design architecture, including NPC identity construction, conditional triggering, information locking, character are progression, and dual prompt engineering frameworks. Despite their formal divergence that one emphasising deliberate multimodal composition and the other immersive interactive narrative, both pathways converge on a shared principle: GenAI is pedagogically most effective when it amplifies, rather than supplants, human critical agency and creative ownership.

读书使人进步，无暇也要读书

‍

欢迎关注我们，一站式获取海量读书资源