Liang He

Liang He

Auto-generated profile, based on publicly available information

Liang He

Microsoft Research AI for Science, Beijing, China

Research Focus & Metrics

Academic Career Length

9Years (2018-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
18 November 2025 in Nature Communications
Verified
Journal Article
Open access
WYWeijie Yin
Weijie Yin
ZZZhaoyu Zhang
Zhaoyu Zhang
SZShuo Zhang
Shuo Zhang
LHLiang He
Liang He
RZRuiyang Zhang
Ruiyang Zhang
ZXZhen Xie
Zhen Xie
Existing RNA language models (RLMs) largely overlook structural information in RNA sequences, leading to incomplete feature extraction and suboptimal performance on downstream tasks. In this study, we present ERNIE-RNA (Enhanced Representations with Base-Pairing Restriction for RNA Modeling), an RNA pre-trained language model based on a modified BERT (Bidirectional Encoder Representations from Transformers). Notably, ERNIE-RNA's attention maps exhibit superior ability to capture RNA structural features through zero-shot prediction, outperforming conventional methods like RNAfold and RNAstructure, suggesting that ERNIE-RNA naturally develops comprehensive representations of RNA architecture during pre-training. Moreover, after fine-tuning, ERNIE-RNA achieves state-of-the-art (SOTA) performance across various downstream tasks, including RNA structure and function predictions. In summary, ERNIE-RNA provides versatile features that can be effectively applied to a wide range of research tasks. Our findings highlight that integrating key knowledge-based priors into the BERT framework may enhance the performance of other language models.

17 March 2024 in bioRxiv
Verified
Preprint
Open access
PDF file
WYWeijie Yin
Weijie Yin
ZZZhaoyu Zhang
Zhaoyu Zhang
LHLiang He
Liang He
RJRui Jiang
Rui Jiang
SZShuo Zhang
Shuo Zhang
ZXZhen Xie
Zhen Xie
With large amounts of unlabeled RNA sequences data produced by high-throughput sequencing technologies, pre-trained RNA language models have been developed to estimate semantic space of RNA molecules, which facilities the understanding of grammar of RNA language. However, existing RNA language models overlook the impact of structure when modeling the RNA semantic space, resulting in incomplete feature extraction and suboptimal performance across various downstream tasks. In this study, we developed a RNA pre-trained language model named ERNIE-RNA ( E nhanced R eprese n tations with base-pa i ring r e striction for RNA modeling) based on a modified BERT (Bidirectional Encoder Representations from Transformers) by incorporating base-pairing restriction with no MSA (Multiple Sequence Alignment) information. We found that the attention maps from ERNIE-RNA with no fine-tuning are able to capture RNA structure in the zero-shot experiment more precisely than conventional methods such as fine-tuned RNAfold and RNAstructure, suggesting that the ERNIE-RNA can provide comprehensive RNA structural representations. Furthermore, ERNIE-RNA achieved SOTA (state-of-the-art) performance after fine-tuning for various downstream tasks, including RNA structural and functional predictions. In summary, our ERNIE-RNA model provides general features which can be widely and effectively applied in various subsequent research tasks. Our results indicate that introducing key knowledge-based prior information in the BERT framework may be a useful strategy to enhance the performance of other language models.

8 September 2025 in Nature Machine Intelligence
Verified
Journal Article
Closed access
PDF file
HSHaoran Sun
Haoran Sun
LHLiang He
Liang He
PDPan Deng
Pan Deng
GLGuoqing Liu
Guoqing Liu
ZZZhiyu Zhao
Zhiyu Zhao
TLTie-Yan Liu
Tie-Yan Liu

4 August 2023
Verified
Conference Paper
Open access
KGKaiyuan Gao
Kaiyuan Gao
LWLijun Wu
Lijun Wu
JZJinhua Zhu
Jinhua Zhu
TPTianbo Peng
Tianbo Peng
YXYingce Xia
Yingce Xia
TLTie-Yan Liu
Tie-Yan Liu
Antibodies are proteins that effectively protect the human body by binding to pathogens. Recently, deep learning-based computational antibody design has attracted popular attention since it automatically mines the antibody patterns from data that could be complementary to human experiences. However, the computational methods heavily rely on high-quality antibody structure data, which is quite limited. Besides, the complementarity-determining region (CDR), which is the key component of an antibody that determines the specificity and binding affinity, is highly variable and hard to predict. Therefore, the limited availability of high-quality antibody structure data exacerbates the difficulty of CDR generation. Fortunately, there is a large amount of sequence data for antibodies that can help model the CDR and reduce reliance on structure data. By witnessing the success of pre-training models for protein modeling, in this paper, we develop the antibody pre-training language model and incorporate it into the antigen-specific antibody design model in a systemic way. Specifically, we first pre-train a novel antibody language model based on the sequence data, then propose a one-shot way for sequence and structure generation of CDR to mitigate the high cost and error propagation associated with autoregressive methods, and finally leverage the pre-trained antibody model for the antigen-specific antibody generation model with some carefully designed modules. Our experiments demonstrate the superiority of our method over previous baselines in tasks such as sequence and structure generation, CDR-H3 design for antigen binding, and antibody optimization1. The code is available at https://github.com/KyGao/ABGNN.

12 February 2018 in IEEE Transactions on Big Data
Verified
Journal Article
Closed access
LHLiang He
Liang He
BSBin Shao
Bin Shao
YXYanghua Xiao
Yanghua Xiao
YLYatao Li
Yatao Li
TLTie-Yan Liu
Tie-Yan Liu
HXHuanhuan Xia
Huanhuan Xia
In this big data era, knowledge becomes increasingly linked, along with the rapid growth in data volume. Connected knowledge is naturally represented and stored as knowledge graphs, which are of more and more importance for many frontier research areas such as machine intelligence. Effectively finding relations between entities in a large knowledge graph plays a key role in many knowledge graph applications, as the most valuable part of a knowledge graph is its rich connectedness, which captures rich information about the objects in the real world. However, due to the intrinsic complexity of real-world knowledge, finding semantically close relations by navigation in a large knowledge graph is very challenging. Canonical graph exploration methods inevitably result in combinatorial explosion especially when the paths connecting two entities are long: the search space is $O(d^l)$ , where $d$ is the average graph node degree and $l$ is the length of the path. In this paper, we will systematically study the semantic navigation problem for large knowledge graphs. Inspired by AlphaGo, which was overwhelmingly successful in the game Go, we designed an efficient semantic navigation method based on a well-tailored Monte Carlo Tree Search algorithm with the unique characteristics of knowledge graphs considered. Extensive experiments on different real-life knowledge bases show that our method is not only effective but also very efficient.

16 November 2023 in bioRxiv
Verified
Preprint
Open access
PDF file
HSHaoran Sun
Haoran Sun
LHLiang He
Liang He
PDPan Deng
Pan Deng
GLGuoqing Liu
Guoqing Liu
ZZZhiyu Zhao
Zhiyu Zhao
TLTie-Yan Liu
Tie-Yan Liu
Protein engineering holds significant promise for designing proteins with customized functions, yet the vast landscape of potential mutations versus limited lab capacity constrains the discovery of optimal sequences. To address this, we present the µ Protein framework, which accelerates protein engineering by combining µ Former, a deep learning model for accurate mutational effect prediction, with µ Search, a reinforcement learning algorithm designed to efficiently navigate the protein fitness landscape using µ Former as an oracle. µ Protein leverages single mutation data to predict optimal sequences with complex, multi-amino acid mutations through its modeling of epistatic interactions and a multi-step search strategy. Except from state-of-the-art performance on benchmark datasets, µ Protein identified high-gain-of-function multi-point mutants for the enzyme β -lactamase, surpassing the highest known activity level, in wet-lab, trained solely on single mutation data. These results demonstrate µ Protein’s capability to discover impactful mutations across vast protein sequence space, offering a robust, efficient approach for protein optimization.