Large Knowledge Model:Perspectives and Challenges

Humankind's understanding of the world is fundamentally linked to our perception and cognition,with human languages serving as one of the major carriers of world knowledge.In this vein,Large Language Models(LLMs)like ChatGPT epitomize the pre-training of extensive,sequence-based world knowledge into neural networks,facilitating the processing and manipulation of this knowledge in a parametric space.This article explores large models through the lens of"knowledge".We initially investigate the role of symbolic knowledge such as Knowledge Graphs(KGs)in enhancing LLMs,covering aspects like knowledge-augmented language model,structure-inducing pretraining,knowledgeable prompts,structured CoT,knowledge editing,semantic tools for LLM and knowledgeable Al agents.Subsequently,we examine how LLMs can boost traditional symbolic knowledge bases,encompassing aspects like using LLM as KG builder and controller,structured knowledge pretraining,and LLM-enhanced symbolic reasoning.Considering the intricate nature of human knowledge,we advocate for the creation of Large Knowledge Models(LKM),specifically engineered to manage diversified spectrum of knowledge structures.This promising undertaking would entail several key challenges,such as disentangling knowledge base from language models,cognitive alignment with human knowledge,integration of perception and cognition,and building large commonsense models for interacting with physical world,among others.We finally propose a five-"A"principle to distinguish the concept of LKM.

New avenues for systematically inferring cellcell communication:through single-cell transcriptomics data

For multicellular organisms,cell-cell communication is essential to numerous biological processes.Drawing upon the latest development of single-cell RNA-sequencing(scRNA-seq),high-resolution transcriptomic data have deepened our understanding of cellular phenotype heterogeneity and composition of complex tissues,which enables systematic cell-cell communication studies at a single-cell level.We first summarize a common workflow of cell-cell communication study using scRNA-seq data,which often includes data preparation,construction of communication networks,and result validation.Two common strategies taken to uncover cell-cell communications are reviewed,e.g.,physically vicinal structure-based and ligand-receptor interaction-based one.To conclude,challenges and current applications of cell-cell communication studies at a single-cell resolution are discussed in details and future perspectives are proposed.

OpenKG Chain:A Blockchain Infrastructure for Open Knowledge Graphs

The early concept of knowledge graph originates from the idea of the semantic Web,which aims at using structured graphs to model the knowledge of the world and record the relationships that exist between things.Currently publishing knowledge bases as open data on the Web has gained significant attention.In China,Chinese Information Processing Society of China(CIPS)launched the OpenKG in 2015 to foster the development of Chinese Open Knowledge Graphs.Unlike existing open knowledge-based programs,OpenKG chain is envisioned as a blockchain-based open knowledge infrastructure.This article introduces the first attempt at the implementation of sharing knowledge graphs on OpenKG chain,a blockchain-based trust network.We have completed the test of the underlying blockchain platform,and the on-chain test of OpenKG’s data set and tool set sharing as well as fine-grained knowledge crowdsourcing at the triple level.We have also proposed novel definitions:K-Point and OpenKG Token,which can be considered to be a measurement of knowledge value and user value.1,033 knowledge contributors have been involved in two months of testing on the blockchain,and the cumulative number of on-chain recordings triggered by real knowledge consumers has reached 550,000 with an average daily peak value of more than 10,000.For the first time,we have tested and realized on-chain sharing of knowledge at entity/triple granularity level.At present,all operations on the data sets and tool sets at OpenKG.CN,as well as the triplets at OpenBase,are recorded on the chain,and corresponding value will also be generated and assigned in a trusted mode.Via this effort,OpenKG chain looks forward to providing a more credible and traceable knowledge-sharing platform for the knowledge graph community.

Neural symbolic reasoning with knowledge graphs:Knowledge extraction,relational reasoning,and inconsistency checking

Knowledge graphs(KGs)express relationships between entity pairs,and many real-life problems can be formulated as knowledge graph reasoning(KGR).Conventional approaches to KGR have achieved promising performance but still have some drawbacks.On the one hand,most KGR methods focus only on one phase of the KG lifecycle,such as KG completion or refinement,while ignoring reasoning over other stages,such as KG extraction.On the other hand,traditional KGR methods,broadly categorized as symbolic and neural,are unable to balance both scalability and interpretability.To resolve these two problems,we take a more comprehensive perspective of KGR with regard to the whole KG lifecycle,including KG extraction,completion,and refinement,which correspond to three subtasks:knowledge extraction,relational reasoning,and inconsistency checking.In addition,we propose the implementation of KGR using a novel neural symbolic framework,with regard to both scalability and interpretability.Experimental results demonstrate that our proposed methods outperform traditional neural symbolic models.

Nanoparticle coagulation and dispersion in a rotating curved pipe

Nanoparticulate flow in a curved pipe differs to that in a straight pipe, and is of scientific and technological interest. The length of flow path within a curved pipe is generally short, and the fluid experiences a centrifugal force which significantly changes the particle distribution both inside and downstream of the pipe. This phenomenon is more obvious when particle coagulation occurs. Prior to this study, nanoparticle migration in a rotating curved pipe with consideration of coagulation and dispersion has not been reported in detail. This study aimed

Selectivity mechanism of Anoplophora glabripennis on four different species of maples

Anoplophora glabripennis(Motsch.)is a wood-boring beetle that is native to China.For a long time,it caused great losses in the economy and ecology of northwest China.Attractants are often used to control insects.The volatiles emitted from the host plant play an important role for insects in finding their target.To explore the mechanism of selec-tivity to different host plants,the response of Anoplophora glabripennis to four different host plants was investigated,which included Acer negundo L.,Acer mono Maxim.,Acer truncatum Bunge.and Acer platanoides L.,and the com-pounds in the profiles of volatiles were identified from these species.The olfactory responses of Anoplophora glabripennis to the odors of different plants showed preference for certain host plants:Acer negundo,Acer mono and Acer truncatum.The attraction of Acer negundo and Acer mono was signifi-cantly different(p<0.05).The attraction of Acer negundo to the insects was stronger than to Acer mono.Acer platanoides neither attracted nor repelled the insects.Compounds in the profiles of volatiles from the above four species were identified and quantified by gas chromatography-mass spec-trometry(GC-MS)equipped with a CP-4020 termodesorp-tion and cold trap(TCT)device.The constituents of volatiles and the relative concentrations were different in the four host species.The amount of ketones,alcohols and aldehydes in the four plants showed the same order:Acer negundo<Acer mono<Acer truncatum<Acer platanoides,while that of alkanes and esters was quiet different:Acer negundo L.>Acer mono Maxim.>Acer truncatum Bunge.>Acer platanoides L..1-penten-3-ol,ocimene and trans-Germanylacetone were repellent to Anoplophora glabripennis.1-penten-3-ol and trans-gerranylacetone were identified in Acer platanoides,and Ocimene was the most attractive to Anoplophora glabripennis among these species.The extent of feeding damage caused by Anoplophora glabripennis differed among four species.The sequences was Acer negundo>Acer mono>Acer truncatum>Acer platanoides.The epidermal hairs of the four host plants revealed that the extent of damage was related to the physical characteristics of the host plants.

Manipulation of Coherent Optical Propagation Based on Monolayer MoS2 Resonator

Atomically thin two-dimensional semiconductor nanomaterials have attained considerable attention currently.We here theoretically investigate the phenomena of slow and superluminal light based on the MoS2 resonator system driven by two-tone fields.Superluminal and ultraslow probe light without absorption can be obtained via manipulating the pump laser on-and off-resonant with the exciton frequency under different parameters regimes,respectively,of which the magnitude is larger than that in a carbon nanotube resonator.The bandwidth of the probe spectrum determined by the quality factor Q of MoS2 resonator is also presented.Furthermore,we also demonstrate the phenomenon of phonon induced transparency and show an optical transistor in the system.The all-optical device based on MoS2 resonator may indicate potential chip-scale applications in quantum information with the currently popular pump-probe technology.