Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Concept Videos

Reaction Quotient02:35

Reaction Quotient

The status of a reversible reaction is conveniently assessed by evaluating its reaction quotient (Q). For a reversible reaction described by m A + n B ⇌ x C + y D, the reaction quotient is derived directly from the stoichiometry of the balanced equation as
Predicting Reaction Outcomes02:24

Predicting Reaction Outcomes

Kinetics describes the rate and path by which a reaction occurs. In contrast, thermodynamics deals with state functions and describes the properties, behavior, and components of a system. It is not concerned with the path taken by the process and cannot address the rate at which a reaction occurs. Although it does provide information about what can happen during a reaction process, it does not describe the detailed steps of what appears on an atomic or a molecular level. On the other hand,...
Measuring Reaction Rates03:09

Measuring Reaction Rates

Polarimetry finds application in chemical kinetics to measure the concentration and reaction kinetics of optically active substances during a chemical reaction. Optically active substances have the capability of rotating the plane of polarization of linearly polarized light passing through them—a feature called optical rotation. Optical activity is attributed to the molecular structure of substances. Normal monochromatic light is unpolarized and possesses oscillations of the electrical field in...
Multi-Step Reactions02:31

Multi-Step Reactions

Chemical reactions often occur in a stepwise fashion involving two or more distinct reactions taking place in a sequence. A balanced equation indicates the reacting species and the product species, but it reveals no details about how the reaction occurs at the molecular level. The reaction mechanism (or reaction path) provides details regarding the precise, step-by-step process by which a reaction occurs. Each of the steps in a reaction mechanism is called an elementary reaction. These...
Quantitative Aspects of Drug-Receptor Interaction01:30

Quantitative Aspects of Drug-Receptor Interaction

The receptor occupancy theory connects a drug's response to the number of occupied receptors. With higher drug concentrations, more receptors are occupied, leading to increased responses. The formation of drug-receptor complexes involves association and dissociation rates, which reach equilibrium when the forward and backward reactions are equal. The equilibrium association constant (Ka) and its inverse, the equilibrium dissociation constant (Kd), indicate drug affinity. Higher Ka and lower Kd...
The Equilibrium Binding Constant and Binding Strength02:18

The Equilibrium Binding Constant and Binding Strength

The equilibrium binding constant (Kb) quantifies the strength of a protein-ligand interaction. Kb can be calculated as follows when the reaction is at equilibrium:

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Self-face recognition under self-implicating threat: preserved self-prioritization and recalibrated control dynamics.

Cognitive research: principles and implications·2026
Same author

Multi-path neural mechanisms of self-other reward integration impairment in schizophrenia.

Neuroscience and biobehavioral reviews·2026
Same author

NMR-Solver: automated structure elucidation via large-scale spectral matching and physics-guided fragment optimization.

Nature communications·2026
Same author

Toward Generalizable Data-Driven Pharmacokinetics with Interpretable Neural ODEs.

Journal of chemical information and modeling·2026
Same author

NMRexp: A database of 3.3 million experimental NMR spectra.

Scientific data·2025
Same author

CD93: A Promising NETs-Related Biomarker for Diagnosis and Therapy in Actinic Keratosis.

OncoTargets and therapy·2025
Same journal

DeepDPM: A Deep Learning Method for MoRFs Prediction Based on Wavelet Transform and Dynamic Convolutional Attention Mechanism.

Journal of chemical information and modeling·2026
Same journal

Graph-Based Generation and Reduction of Complex Chemical Reaction Networks.

Journal of chemical information and modeling·2026
Same journal

Modeling the Sensitivity of Large-Scale Virtual Screening to Scoring Function Accuracy, Artifacts, and Library Composition.

Journal of chemical information and modeling·2026
Same journal

Machine Learning-Driven Discovery of Indole/Oxoindole-Piperazine Scaffolds as Dual MAO-B/Sig-1R Ligands for Neurodegenerative Disorders.

Journal of chemical information and modeling·2026
Same journal

Mapping Evolution of Molecules across Biochemistry with Assembly Theory.

Journal of chemical information and modeling·2026
Same journal

Structural Proteomics-Based Deciphering of Hydrophobic Packing Fingerprints Informing Protein Thermostability in TIM Barrels.

Journal of chemical information and modeling·2026
See all related articles

Related Experiment Video

Updated: May 16, 2026

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness
03:14

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness

Published on: December 6, 2024

RxnBench: A Multimodal Benchmark for Evaluating Large Language Models on Chemical Reaction Understanding from the

Hanzheng Li1,2, Xi Fang1, Yixuan Li3

  • 1DP Technology, Beijing 100080, China.

Journal of Chemical Information and Modeling
|May 14, 2026
PubMed
Summary
This summary is machine-generated.

Multimodal large language models (MLLMs) struggle with chemical reaction understanding from scientific PDFs. A new benchmark, RxnBench, reveals significant gaps in visual and logical reasoning, highlighting the need for specialized AI chemists.

More Related Videos

A Bilingual Computational Workflow for Identifying Potential PLK1 Inhibitors in American Sign Language and English
14:34

A Bilingual Computational Workflow for Identifying Potential PLK1 Inhibitors in American Sign Language and English

Published on: April 3, 2026

Related Experiment Videos

Last Updated: May 16, 2026

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness
03:14

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness

Published on: December 6, 2024

A Bilingual Computational Workflow for Identifying Potential PLK1 Inhibitors in American Sign Language and English
14:34

A Bilingual Computational Workflow for Identifying Potential PLK1 Inhibitors in American Sign Language and English

Published on: April 3, 2026

Area of Science:

  • Chemistry
  • Artificial Intelligence
  • Scientific Literature Analysis

Background:

  • Multimodal large language models (MLLMs) show promise for scientific discovery.
  • Evaluating MLLMs' comprehension of chemical reactions in scientific literature is crucial but underexplored.
  • Authentic scientific PDFs present complex graphical and textual information.

Purpose of the Study:

  • Introduce RxnBench, a novel benchmark for evaluating MLLMs' chemical reaction understanding.
  • Assess MLLMs' capabilities in visual perception, mechanistic reasoning, and cross-modal information synthesis from scientific PDFs.
  • Identify current limitations of MLLMs in interpreting complex chemical information.

Main Methods:

  • Developed RxnBench, a multitiered benchmark with two tasks: Single-figure Question Answering (SF-QA) and Full-Document Question Answering (FD-QA).
  • SF-QA includes 1525 questions from 305 curated reaction schemes, testing visual perception and mechanistic reasoning.
  • FD-QA involves synthesizing information from 108 articles, requiring integration of text, schemes, and tables.

Main Results:

  • MLLMs demonstrate proficiency in extracting explicit text but falter in deep chemical logic and structural recognition.
  • Models with inference-time reasoning show improved performance over standard architectures.
  • No evaluated MLLMs achieved 50% accuracy on the challenging Full-Document QA task.

Conclusions:

  • A significant capability gap exists in MLLMs' understanding of chemical reactions from authentic literature.
  • Domain-specific visual encoders and enhanced reasoning engines are essential for advancing AI in chemistry.
  • Further development is needed to create truly autonomous AI chemists capable of complex scientific reasoning.