講座1：
講座主題：A Novel TCM-based AI Large Model Framework toward Human diseases and Drug-Diseases Associations
講座時(shí)間: 2023年10月30日星期一 9:00
講座地點(diǎn): 沙河校區(qū)理學(xué)樓202
主 講 人: 田亮 香港浸會(huì)大學(xué)物理系
摘要:
Traditional Chinese Medicine (TCM), which originated in ancient China with a history of thousands of years, characterizes and addresses human physiology, pathology, and diseases diagnosis and prevention using TCM theories and Chinese herbal products. Recently, the World Health Organization included TCM in the global diagnostic compendium, which marks the international recognition of TCM in global health care. Considering this, many research works have been devoted to revealing the effectiveness and efficacy of Chinese herbs for new drug discovery in a bottom-up manner. However, the pharmacological principles in TCM theory, the core treasure house of TCM, have rarely been systematically investigated in a top-down manner, which hinders the modernization and standardization of TCM. To bridge the gap, we propose a novel TCM-based artificial intelligence (AI) framework to unravel general patterns and principles of human disease and investigate potential drug-diseases associations. We collect and refine extensive TCM data, as well as biological, chemical, and clinical data, to establish an integrated multi-modal TCM database. Subsequently, we construct a TCM pharmacological network to reveals the underlying structure and patterns within the TCM data. An attention-based AI model is trained to embed multi-modal TCM data into an interpretable pharmacological space, allowing for quantitative and personalized analysis of complex interactions among diseases, symptoms, herbs, compounds, and genes. The pharmacological embedding space with biological significance provides new perspectives toward modern medicine issues from the view of TCM. Our work aims to promote the quantitative underpinning of TCM pharmacological principles, provide a basis for the objectification of the diagnosis and treatment process of TCM, and pave the way for the knowledge fusion of TCM evidence-based medicine and modern biology.
報(bào)告人簡(jiǎn)介:
田亮博士現(xiàn)為香港浸會(huì)大學(xué)（HKBU）物理系助理教授，并兼任香港浸會(huì)大學(xué)環(huán)境與生物分析國(guó)家重點(diǎn)實(shí)驗(yàn)室和計(jì)算與理論研究所的學(xué)術(shù)成員，以及深圳京魯計(jì)算科學(xué)與應(yīng)用研究院成員。在加入香港浸會(huì)大學(xué)之前，他曾先后在哈佛大學(xué)醫(yī)學(xué)院擔(dān)任博士后研究員，在美國(guó)波士頓布萊根婦女醫(yī)院查寧網(wǎng)絡(luò)醫(yī)學(xué)部擔(dān)任研究員。田博士的團(tuán)隊(duì)利用各種分析、數(shù)值、建模、模擬、數(shù)據(jù)挖掘和機(jī)器學(xué)習(xí)技術(shù)，對(duì)復(fù)雜系統(tǒng)、統(tǒng)計(jì)物理和生物物理學(xué)進(jìn)行前沿的跨學(xué)科研究。主要研究方向包括人類微生物組與群落生態(tài)學(xué)、生物大數(shù)據(jù)與機(jī)器學(xué)習(xí)、復(fù)雜網(wǎng)絡(luò)：結(jié)構(gòu)與動(dòng)力學(xué)、中醫(yī)數(shù)據(jù)挖掘、流行病學(xué)建模等。
講座2：
講座主題：時(shí)序信息加工的神經(jīng)計(jì)算機(jī)制及其類腦算法研究
講座時(shí)間: 2023年10月30日星期一 10:00
講座地點(diǎn): 沙河校區(qū)理學(xué)樓202
主 講 人: 弭元元 清華大學(xué)心理學(xué)系
摘要:
Temporal sequence processing is fundamental in brain cognitive functions. Experimental data has indicated that the representations of ordinal information and contents of temporal sequences are disentangled in the brain, but the neural mechanism underlying this disentanglement remains largely unclear. We investigate how recurrent neural circuits learn to represent the abstract order structure of temporal sequences, and how the disentangled representation of order structure facilitates the processing of temporal sequences. We show that with an appropriate training protocol, a recurrent neural circuit can learn tree-structured attractor dynamics to encode the corresponding tree-structured orders of temporal sequences. This abstract temporal order template can then be bound with different contents, allowing for flexible and robust temporal sequence processing. Using a transfer learning task, we demonstrate that the reuse of a temporal order template facilitates the acquisition of new temporal sequences, if these sequences share the same or partial ordinal structure. Using a key-word spotting task, we demonstrate that the tree-structured attractor dynamics improves the robustness of temporal sequence discrimination, if the ordinal information is the key to differentiate these sequences.
報(bào)告人簡(jiǎn)介:
弭元元，清華大學(xué)心理學(xué)系副教授，研究方向?yàn)橛?jì)算神經(jīng)科學(xué)。專注于研究腦在網(wǎng)絡(luò)層面處理動(dòng)態(tài)信息的一般性原理，包括工作記憶的容量與調(diào)控、時(shí)空信息的網(wǎng)絡(luò)編碼等；基于此發(fā)展了類腦運(yùn)動(dòng)模式的快速識(shí)別算法、運(yùn)動(dòng)目標(biāo)的預(yù)測(cè)追蹤算法等，并與工業(yè)界合作探索這些類腦算法在實(shí)際場(chǎng)景中的應(yīng)用。以第一或者通訊（含共同）在神經(jīng)科學(xué)頂級(jí)期刊Neuron, PNAS, Progress in Neurobiology，人工智能頂會(huì)NeurIPS等發(fā)表論文20余篇。獲得國(guó)家自然科學(xué)基金委交叉學(xué)部?jī)?yōu)秀青年基金和北京市科技新星計(jì)劃等項(xiàng)目的支持。
