The team targeted a domain where the need for agile evidence synthesis is particularly acute: randomized controlled trials (RCTs) that rely on digital health technologies. From wearable activity trackers and smartphone-based interventions to telehealth platforms and remote monitoring sensors, digital tools have multiplied the ways clinical trials collect and report outcomes. That proliferation has outpaced the reporting guidelines designed to standardize trial publication, most notably CONSORT, the widely adopted checklist that establishes common data elements for trial reporting. When researchers attempt to pool results across digital health trials, they routinely encounter substantial heterogeneity in how outcomes are described, measured, and tabulated.

This reporting heterogeneity is more than an inconvenience. Systematic reviews depend on the ability to locate and compare consistent data elements across dozens or hundreds of primary studies. When a trial reports its engagement metrics in one format and its equity-related participant characteristics in another, or omits key technology-adherence details altogether, reviewers must either exclude the study or spend additional time hunting through supplementary materials. The authors argue that machine-assisted evaluation of reporting adherence offers a pathway toward more nimble information management, one that can accommodate the diverse and disruptive nature of technology-enabled trials as they emerge.

The framework the researchers propose follows a two-stage logic that mirrors how human reviewers work, but compresses the timeline dramatically. In the retrieval stage, the large language model performs article screening, deciding which candidate publications meet inclusion criteria for a given review question. In the abstraction stage, the model handles text and table preprocessing and then conducts full-text review, extracting the specific data elements the reviewers need. Crucially, the approach relies on an unmodified pretrained model, meaning the researchers did not train or fine-tune the model on annotated medical literature. Instead, they designed prompts and workflows that steer the general-purpose model toward reliable performance on specialized tasks.

To test the framework, the team ran two case studies, each examining adherence to a different set of reporting guidelines in digital health technology-enabled RCTs. The first case study evaluated how well trials reported technology-related outcomes, such as measures of participant engagement with the digital intervention. The second focused on digital health equity reporting items, which capture whether trials describe how their technologies perform across diverse populations, including groups that have historically been underrepresented in digital health research. Both case studies were chosen because they represent areas where reporting practices remain inconsistent and where automated assessment could deliver immediate practical value.

The performance results are the study’s headline finding. Across both screening and information-extraction tasks, the large language model achieved F1 scores of at least 0.9, a level the authors describe as on par with inter-annotator agreement, the benchmark that measures how well two trained human reviewers agree with each other on the same tasks. In practical terms, this means the model’s judgments about which articles to include and its extractions of specific data points from full texts were as consistent with a reference standard as a second human expert would typically be. For a tool that requires no custom training, that level of agreement represents a meaningful threshold for real-world deployment.

The implications extend well beyond the two case studies. Systematic reviews of RCTs underpin clinical practice guidelines, regulatory decisions, and health technology assessments, yet the volume of published trials continues to grow faster than the workforce available to review them. A framework that can reliably automate screening and abstraction could shorten the lag between the publication of primary evidence and its incorporation into synthesized knowledge. In digital health specifically, where interventions and technologies evolve on timescales of months rather than years, that acceleration could determine whether reviews remain relevant by the time they are published.

The study also carries a pointed message for trial investigators and journal editors. The authors demonstrate the importance of common reporting elements across digital health technology-enabled RCTs, and their automated assessment of reporting adherence makes gaps visible at scale. If a machine can rapidly quantify how consistently trials report technology engagement or equity-related items, then funders, guideline developers, and editorial boards gain a new instrument for monitoring compliance and identifying where CONSORT extensions or digital-health-specific checklists need strengthening. Better reporting, in turn, feeds back into better synthesis, creating a virtuous cycle that the framework is positioned to support.

Methodologically, the retrieval-to-abstraction design reflects a broader shift in how researchers are approaching large language models in scientific workflows. Rather than building bespoke models for each task, the authors show that a single pretrained model, guided by carefully structured prompts and pipeline stages, can handle the heterogeneous formats found in real published literature, including the tables and supplementary text where trial outcomes often hide. The preprocessing steps that convert raw text and tables into model-readable inputs are a key part of the pipeline, addressing one of the persistent obstacles to automating literature review: the sheer variability of how information is presented across journals and study designs.

Cautious optimism seems the appropriate stance. The study reports case demonstrations in a specific domain rather than a comprehensive validation across all of medicine, and the authors frame their findings as supporting the scalability and rigor of accelerated evidence synthesis, not as proof that human reviewers are obsolete. Human expertise remains essential for defining review questions, adjudicating ambiguous cases, and interpreting clinical significance. But the demonstration that an unmodified large language model can match inter-annotator agreement on both screening and extraction in digital health trial reviews marks a tangible step toward a future in which the evidence base for clinical decisions is updated as quickly as the technologies it evaluates. As digital health continues to reshape how trials are run, the tools for reading those trials may finally be catching up.

Subject of Research: Large language model-assisted literature synthesis for evaluating reporting adherence in digital health technology-enabled randomized controlled trials

Article Title: Retrieval-to-abstraction LLM-assisted literature synthesis: Case demonstrations in digital health technology-enabled randomized controlled trial outcome reporting

Article References: Harrison, T. B., Hu, D., Jia, H., Tan, S. H., Chen, F., Zhang, Z., Lu, Q., Wang, J., Wang, L., Huang, M., Prokop, L. J., Hoy, M., St. Sauver, J., Fu, S., & Liu, H. (2026). Retrieval-to-abstraction LLM-assisted literature synthesis: Case demonstrations in digital health technology-enabled randomized controlled trial outcome reporting. PLOS Digital Health, 5(10), e0001768. https://doi.org/10.1371/journal.pdig.0001768

Image Credits: AI Generated

DOI: 10.1371/journal.pdig.0001768

Keywords: large language models, systematic reviews, randomized controlled trials, digital health, CONSORT reporting guidelines, evidence synthesis, natural language processing, clinical trial reporting, digital health equity, information extraction, article screening, PLOS Digital Health

The study, published in Earth Science Informatics, was led by Ke Li of the Chongqing AI Geological and Mineral Research Institute together with colleagues from the Chongqing Bureau of Geology and Minerals Exploration and partner institutions. Their starting point was a problem familiar to anyone who has worked in applied geoscience: the knowledge needed to assess a dangerous rock slope is scattered across heterogeneous documents, including field investigation reports, engineering records and monitoring archives, each organized differently and none designed for machine reading. This fragmentation complicates three things at once, the authors note: modeling the chains of events that turn a trigger such as a storm into a disaster, extracting reliable facts from messy text, and retrieving both specific entity-level details and broader mechanism-level context when a question is asked.

To solve it, the team built what they call an ontology-guided LightRAG framework, a system with three interlocking components. The first is a domain ontology, a formal specification of the concepts and relationships that matter in precarious rock hazard work. Ontologies have a long pedigree in geoinformatics, tracing back to foundational work on semantic web terminology for earth and environmental science, and they serve as a shared vocabulary that both humans and machines can use. The researchers’ ontology contains 31 entity categories organized across five core evolutionary stages: environmental background, triggering factors, disaster evolutionary processes, mitigation decisions and effectiveness evaluation. In effect, it encodes how geologists think about a rockfall hazard from first causes to final remediation.

The second component is a four-stage conversational extraction process driven by GeoGPT, an open-source geological language model fine-tuned from Qwen2.5-72B-Instruct by its original developers and deployed locally by the team without further modification. In the first stage, the system performs an ontology-aware semantic scan of a document, using a sliding window to pull in surrounding paragraphs so that cross-sentence context is preserved. In the second stage, it anchors entities and resolves co-references, mapping phrases such as ‘this rock mass’ or a shorthand label like ‘W01’ to a single unique identifier. The third stage involves recursive attribute backfilling, in which the model hunts through the text for missing physical parameters such as volumes, fracture orientations and stability coefficients, and the fourth applies logical verification, checking that numbers and qualitative descriptions agree before stripping away redundant modifiers and emitting structured data.

The authors illustrate the pipeline with a real example from Chaqi Mountain, where a precarious rock unit designated W01 sits on a second-tier cliff above a school housing roughly 4,000 students and staff. From a single passage, the system extracts the unit’s irregular geometry and volume of about 66.4 cubic meters, its dominant sliding-falling failure mode, the controlling fracture set with its attitude expressed as strike and dip angles, the weak Lower Jurassic mudstone pedestal that softens and forms cavities under weathering, and three stability coefficients, 8.62 under natural conditions, 2.00 under storm conditions and 5.48 under seismic conditions, all confirming a currently stable state. That kind of multi-fact extraction, performed consistently across an entire archive, is precisely what manual digitization struggles to deliver.

The third component is the retrieval architecture. The team adopted LightRAG, a graph-based retrieval-augmented generation approach, and paired it with a hybrid document retrieval strategy. Questions and document chunks are encoded with the BGE-M3 embedding model into 1,024-dimensional dense vectors, while a classic BM25 keyword retriever runs in parallel; the two ranked lists are merged using Reciprocal Rank Fusion with a constant of 60, and the top ten chunks are retained. On the graph side, entities in a question are linked to knowledge-graph nodes by exact, alias or semantic matching, then the system traverses incoming and outgoing relations up to two hops, keeping only simple paths without repeated nodes. The ten best paths, ranked by cosine similarity between question and path embeddings, are serialized as head-relation-tail triplets and handed to the language model alongside the document context. This dual retrieval lets the system answer both pinpoint factual queries and broader questions about failure mechanisms.

The performance numbers are striking. With the complete extraction configuration, the system achieved a precision of 89.00 percent, a recall of 87.74 percent and an F1-score of 88.36 percent against a ground-truth dataset annotated by five professor-level senior and senior engineers, each with more than ten years of experience in engineering geology. Ablation analysis showed that the four-stage conversational extraction process was the primary driver of that performance, with GeoGPT itself contributing a smaller additional benefit under the same extraction setting. The annotation protocol itself was rigorous: two experts independently labeled 1,200 text chunks, from which 212 entities were drawn, reaching an initial agreement rate of 91.8 percent and a Cohen’s Kappa of 0.894, with a third senior expert adjudicating all disputes. Most disagreements, 73.5 percent of them, clustered at the boundary between environmental background and triggering factors, a reminder that even expert judgment finds some geological distinctions genuinely ambiguous.

For question answering, the team ran a blinded pairwise evaluation across four precarious rock task categories, comparing GeoGPT against two powerful general-purpose models, DeepSeek-V3 accessed through its official API and Qwen2.5-72B-Instruct deployed locally on eight A100 GPUs with BF16 precision and greedy decoding at temperature zero. The GeoGPT-based system achieved higher overall pairwise win rates than both competitors across all four categories, with the largest margins appearing in the most demanding tasks: stability assessment and mitigation planning. A qualitative example shows why. Asked to analyze collapse mechanisms, DeepSeek-V3 assembled a competent descriptive summary of physical parameters, while GeoGPT structured its answer along explicit geological causal chains, incorporating formal failure mode classifications such as sliding, toppling and composite modes, and even geomechanical expressions like the toppling coefficient equation, aligning its output with domain standards for hazard assessment.

The implications reach well beyond one mountain range. Landslides and rockfalls are among the deadliest geological hazards worldwide, and climate change, slope urbanization and aging infrastructure are expected to increase exposure in many regions. Knowledge graphs have already proven their value in geoscience for link prediction and disaster cause analysis, and large language models are being applied to everything from landslide image analysis to earthquake fatality estimation. What this study adds is a template for combining them responsibly: an ontology to constrain what the machine can extract, a staged conversational process to keep extraction faithful to the source text, and graph-aware retrieval to ground answers in verified facts rather than the model’s own tendencies toward hallucination, a known weakness of generative systems.

There are honest limits. The datasets analyzed in the study cannot be released publicly because of multi-party data-sharing agreements and confidentiality restrictions involving the regional geological survey institutions of the Three Gorges Project, though the authors say data are available upon reasonable request with formal permission from the competent geological authorities. The evaluation also rests on a specific corpus and a specific panel of expert annotators, so the reported win rates describe performance within that setting rather than a universal guarantee. Even so, the framework demonstrates something quietly transformative: that the tacit expertise embedded in thousands of pages of engineering reports can be formalized, verified and made conversational. For the geologists watching rock slopes above schools and highways, an assistant that answers mechanism-level questions in seconds, grounded in their own field’s ontology, could change how hazard assessments get done.

Subject of Research: An ontology-guided LightRAG framework for knowledge-graph construction and question answering on precarious rock hazards

Article Title: An ontology-guided LightRAG framework for knowledge-graph construction and question answering on precarious rock hazards

Article References: Li, K., Pan, Y., Yi, Z., Guo, S., Liu, P., Dong, L., Wang, L., Wang, Y., Qin, W., Cheng, X., She, N., & Yuan, J. (2026). An ontology-guided LightRAG framework for knowledge-graph construction and question answering on precarious rock hazards. Earth Science Informatics, 19(11), Article 202. https://doi.org/10.1007/s12145-026-02258-9

Image Credits: AI Generated

DOI: 10.1007/s12145-026-02258-9

Keywords: precarious rock masses, knowledge graph, ontology, LightRAG, retrieval-augmented generation, GeoGPT, large language models, geohazard assessment, Three Gorges, information extraction, question answering, landslide hazards

To understand why ALTAI matters, it helps to grasp the peculiar challenge that Chinese poses to machines. In English, a named entity such as a company or a person is usually delimited by spaces and signaled by an initial capital letter, so a model can often spot boundaries with relative ease. Chinese text offers no such signposts. The same character sequence can be segmented into words in several plausible ways, and the meaning of a character often depends on which words it belongs to. Consider that a single character might stand alone as a word, join with a neighbor to form a two-character word, or sit inside a longer multi-character phrase. A model that only looks at characters misses the lexical knowledge embedded in dictionaries, while a model that only looks at words risks losing sight of the fine-grained compositional structure of the language. The best-performing systems therefore try to use both levels of information simultaneously.

For years, the dominant approach has been the so-called word-character lattice framework. The idea is to build a lattice structure over the sentence: characters form the backbone, and every dictionary word that matches a span of characters is attached as an additional node. This allows the model to consult word-level information without committing to a single segmentation. A well-known implementation of this idea is the Flat-Lattice Transformer, which folds lattice nodes into a transformer architecture so that characters and words attend to one another. Lattice-based models have delivered solid gains, and they remain widely used in Chinese NER research. Yet, according to the authors of the new study, these methods share a common blind spot.

That blind spot concerns how lexical information is actually integrated. In existing lattice systems, word-level knowledge is typically injected through a dedicated encoder architecture, and the interactions between character representations and word representations are handled implicitly, as a byproduct of attention mechanisms. The authors argue that this design overlooks the complex interactions across different granularities between character-level and word-level representations. In other words, the model may receive word information, but it does not explicitly learn how that information should reshape its understanding of the individual characters that compose the words. Semantic dependencies, too, are underexploited, particularly in their interactions with character representations. The result is that lexical knowledge, one of the most valuable resources available for Chinese NER, is not fully exploited.

ALTAI addresses this weakness with two tightly coupled components. The first is a Cross Transformer, a module designed to explicitly model interactions among three kinds of representations: the character representations that form the basic units of the text, the word-level lexical representations drawn from matching dictionary entries, and semantic representations that capture higher-level meaning. Rather than letting these views of the sentence mix only incidentally, the Cross Transformer forces a direct exchange of information across the different granularities. Each character can consult the words it participates in, and the semantic layer can modulate how both characters and words are interpreted. This explicit cross-granularity modeling is what distinguishes ALTAI from earlier lattice encoders, where the dialogue between levels of representation was left largely to chance.

The second component is Cross-Granularity Contrastive Learning, a training strategy borrowed from the recent wave of self-supervised learning research. Contrastive learning works by teaching a model to pull related representations together and push unrelated ones apart in the embedding space. In ALTAI, the technique is applied across granularities: character representations are aligned with their corresponding lexical and semantic views, so that the model is encouraged to produce consistent representations across the character, word, and semantic spaces. A character that participates in a meaningful word should end up with an embedding that reflects that word’s presence, and the semantic view of the sentence should agree with what the character-level view suggests. This alignment acts as an additional training signal, shaping the representation space so that entities become more discriminative, meaning that the vectors for genuine entities stand apart more clearly from those of ordinary text.

The combination of these two ideas produces what the authors describe as discriminative entity representations by modeling cross-granularity interactions. The intuition is straightforward: when a character-level representation has been explicitly refined by lexical and semantic information, and when the training process has actively enforced consistency among those views, the final representation of an entity span carries far richer evidence than a representation built from characters alone. Downstream, the model’s sequence labeling decisions, deciding where an entity starts, where it ends, and what type it is, rest on this enriched foundation. The contrastive objective also serves as a form of regularization, discouraging the model from latching onto spurious patterns that appear in one granularity but not in others.

The empirical case for ALTAI rests on experiments across four benchmark datasets for Chinese named entity recognition. The authors report that ALTAI consistently outperforms strong baseline models, including the lattice-based architectures that have defined the state of the art in this area. Consistency across multiple datasets is a meaningful claim in NER research, where a method that excels on one corpus, say one dominated by news text, may falter on another drawn from social media or specialized domains. By holding its advantage across all four benchmarks, ALTAI suggests that its benefits stem from the general principle of explicit cross-granularity interaction rather than from quirks of any particular dataset. The work was supported by the National Natural Science Foundation of China and several Guangdong provincial research programs, reflecting sustained institutional investment in Chinese-language AI research.

The significance of this study extends beyond Chinese NER itself. Named entity recognition is a foundational task that feeds countless downstream applications: information extraction, question answering, knowledge graph construction, search, and recommendation systems all depend on reliably identifying entities in raw text. Improvements at this level propagate through entire pipelines. Moreover, the core insight of ALTAI, that representations at different granularities should be explicitly interacted and aligned rather than merely concatenated or implicitly mixed, is a design principle that could transfer to other languages and tasks where multiple levels of linguistic structure coexist. Languages without clear word boundaries, and tasks such as relation extraction or event detection that similarly straddle character, word, and sentence levels, are natural candidates.

The study also highlights a broader trend in machine learning: the migration of contrastive learning from computer vision into structured language tasks. Originally popularized for learning image representations without labels, contrastive objectives are now being adapted to align representations across views, modalities, and, as here, linguistic granularities. For Chinese NER specifically, the arrival of ALTAI signals a shift in how researchers think about lexical knowledge, not as a static feature to be injected into an encoder, but as a set of interacting perspectives that a model must learn to reconcile. As the field moves forward, the open-access publication of this work, published on 1 September 2026 with a permanent DOI, means that researchers worldwide can examine, reproduce, and build upon the approach, potentially accelerating progress on a problem that sits at the heart of making machines truly literate in one of the world’s most widely spoken languages.

Subject of Research: Cross-granularity contrastive learning for Chinese named entity recognition

Article Title: Semantic-aware cross-granularity contrastive learning for Chinese named entity recognition

Article References: Mao, S., Feng, Z., Li, J., Sun, C., Xiang, D., & Jiang, Y. (2026). Semantic-aware cross-granularity contrastive learning for Chinese named entity recognition. Complex & Intelligent Systems. https://doi.org/10.1007/s40747-026-02483-1

Image Credits: AI Generated

DOI: 10.1007/s40747-026-02483-1

Keywords: Chinese named entity recognition, natural language processing, contrastive learning, Cross Transformer, lattice framework, lexical knowledge, semantic representations, representation learning, benchmark datasets, Complex & Intelligent Systems, deep learning, information extraction

The scale of the problem is easy to underestimate. Curated databases do exist that catalogue which proteins carry which modifications and at which amino acid positions, and these resources are indispensable for experimentalists planning experiments or interpreting mass-spectrometry data. But most of those databases cover only a limited number of PTM types, and they are not updated regularly. Meanwhile, the volume of PTM-related findings buried in PubMed abstracts keeps growing year after year. The result is a widening information gap: knowledge that has already been discovered and published sits locked inside unstructured text, inaccessible to computational analysis, while the manual curation effort needed to unlock it falls further behind. Building a separate, dedicated text-mining system for every one of the hundreds of known PTM types is simply not feasible.

GenPTM attacks this problem with a deliberately generalizable design. The central insight behind the framework is that, despite the bewildering chemical diversity of PTMs, the sentences scientists use to describe them are remarkably similar in structure. A paper might report that a protein is ubiquitinated at a lysine residue, phosphorylated on a serine, or methylated at an arginine, but the underlying linguistic pattern, in which a protein, a modification event, and a site are linked together, repeats across modification types. Rather than teaching a model the vocabulary of each modification separately, the Delaware team replaces PTM-specific modification names and chemical group mentions with generic placeholders. This unified text representation strategy forces the model to focus on the shared textual patterns that express modification events, regardless of which particular modification is being discussed.

On top of this representation, the researchers fine-tuned a classifier built on BiomedBERT, a language model pre-trained on biomedical text. The classifier’s job is deceptively simple: given a candidate protein or a candidate amino acid site in a sentence, decide whether that protein or site is genuinely modified in the context described. This is a harder judgment than it sounds, because scientific prose is full of negations, speculations, and references to other studies. A sentence saying that a site was not found to be phosphorylated, or that phosphorylation at a position is suspected but unconfirmed, must be distinguished from a clean report of an observed modification. The fine-tuned BiomedBERT classifier makes that determination for each candidate, and a post-processing module then assembles the individual decisions into final predictions: modified proteins, modified sites, or complete protein-site pairs.

The training and evaluation strategy is what makes the claim of generalizability convincing. The model was trained on data covering five major PTM types, including well-studied modifications such as ubiquitination and phosphorylation, which dominate the literature and the curated databases. The real test, however, came from evaluation on eight additional PTM types that the model had never seen during training. These included PTMs that are mentioned only rarely in scientific articles, such as citrullination and AMPylation, precisely the kinds of modifications for which dedicated curation resources are weakest and the need for automated extraction is greatest. A system that only works on abundant, well-documented modifications would offer little advantage over existing tools; GenPTM was designed to work on the long tail.

The results reported in the paper are striking. Across all PTM types tested, including the eight held-out modifications, GenPTM achieved F1-scores ranging from 92 percent to 96 percent across three different evaluation categories. The F1-score is a standard metric that balances precision, the fraction of predicted modifications that are correct, against recall, the fraction of true modifications that are found. Scores in the mid-nineties for modification types the model was never explicitly trained on indicate that the placeholder-based representation genuinely transfers knowledge about how modification events are expressed in text. In practical terms, a curator or database builder could point GenPTM at a PTM type with little or no dedicated training data and still expect reliable extraction of modified proteins and their sites from PubMed abstracts.

The implications for proteomics research extend beyond convenience. PTM-agnostic information extraction means that newly discovered or obscure modification types can be surveyed systematically as soon as reports appear, without waiting for a specialized tool or a funded curation project. It also means that existing PTM databases could be updated far more frequently, with automated systems flagging candidate modifications from the newest literature for expert review. Because the framework operates on abstracts, the most accessible layer of the literature, it can scale to the full breadth of PubMed rather than being restricted to a handful of model organisms or heavily studied protein families. The authors position the work as a viable solution for automated PTM knowledge discovery, and the architecture suggests a path toward text-mining systems that generalize across related biomedical relation-extraction problems as well.

The project reflects a substantial collaborative effort. The work was carried out by Shovan Bhowmik, Chuming Chen, Cathy Wu, and K. Vijay-Shanker at the University of Delaware, together with Karen Ross of Georgetown University Medical Center. The team acknowledges the valuable expertise in data curation contributed by Dr. Cecilia Arighi of the University of Delaware, as well as the dedicated curation efforts of Amos Nyabuti and Winnie Aketch, master’s students at Delaware’s Center for Bioinformatics and Computational Biology, who helped develop the training and testing corpora on which the framework depends. High-quality annotated corpora are the quiet foundation of every successful biomedical language model, and the paper is a reminder that advances in artificial intelligence for biology rest on meticulous human annotation work.

Financial support came from multiple National Institutes of Health grants, including R35GM141873, U54GM104941, and P20GM103446, along with 1R24GM146616-01 and, from the NIH Office of Strategic Coordination and the Common Fund, 1OT2OD032092. The article is published open access under a Creative Commons Attribution 4.0 license, meaning that any research group, database curator, or tool developer can read, reuse, and build upon the work without restriction. The paper was received on 30 April 2026, accepted on 14 September 2026, and published on 24 September 2026, carrying the DOI 10.1186/s12859-026-06663-1.

For a field drowning in its own literature, GenPTM offers something rarer than another incremental benchmark improvement: a demonstration that a single, thoughtfully designed system can serve the entire spectrum of post-translational modifications, from the phosphorylation events catalogued in thousands of papers to the citrullination and AMPylation events scattered across a comparative handful. If the framework’s generalization holds as it is applied more broadly, the tedious bottleneck between publication and curated knowledge could begin to close, letting the proteins’ chemical vocabulary be read as fast as scientists can write it.

Subject of Research: Automated information extraction of protein post-translational modifications from scientific literature using a generalizable natural language processing framework

Article Title: GenPTM: a generalizable framework for protein post-translational modification information extraction from the scientific literature

Article References: Bhowmik, S., Ross, K., Chen, C., Wu, C., & Vijay-Shanker, K. (2026). GenPTM: a generalizable framework for protein post-translational modification information extraction from the scientific literature. BMC Bioinformatics. https://doi.org/10.1186/s12859-026-06663-1

Image Credits: AI Generated

DOI: 10.1186/s12859-026-06663-1

Keywords: post-translational modification, information extraction, text mining, BiomedBERT, phosphorylation, ubiquitination, proteomics, PubMed, natural language processing, protein sites, BMC Bioinformatics, machine learning

The research, conducted by Yachuan Zhang and Yi Guo of East China University of Science and Technology in Shanghai, identifies two problems that previous work has largely overlooked. The first concerns the text itself. Social media posts are short, informal, and riddled with abbreviations, slang, emojis, and grammatical irregularities. When a language model tries to extract semantic representations from such input, the intrinsic noise of the writing style makes the resulting meaning representations unreliable. Most earlier studies concentrated on filtering noise from the visual side of the pairing, treating the text as a comparatively stable anchor. Zhang and Guo argue that this assumption is flawed: if the textual foundation is shaky, everything built on top of it inherits the instability. Their first contribution is therefore a Textual Token Refinement network, designed to distill concise, dependable semantic representations from noisy inputs before any cross-modal reasoning takes place.

The second problem concerns how images are used. In many existing systems, visual information enters the model as coarse-grained, global features, essentially a single vector summarizing the whole picture. The authors point out that such coarse features often introduce irrelevant interference rather than useful relational evidence. A photograph accompanying a post may contain dozens of objects, backgrounds, and incidental details that have nothing to do with the two entities whose relationship the system is trying to determine. Feeding the entire image into the fusion process risks drowning the signal in visual clutter. DNMMA addresses this with a multi-granularity visual augmentation strategy that works at several levels of detail simultaneously, combining hierarchical fusion of global features with carefully selected local regions, so that the model can draw on both the overall scene and the small patches that actually carry relational meaning.

One of the more inventive components of the framework is the Adaptive Semantic Mixup module. Data augmentation, the practice of creating synthetic training examples to improve model robustness, is well established in computer vision, but blindly mixing images can produce combinations that contradict the accompanying text. The Adaptive Semantic Mixup balances visual diversity against text-image consistency by adaptively fusing synthetic and original images. In practice, the system decides how much of a generated or mixed image to blend with the original photograph, ensuring that the augmented visuals remain semantically compatible with the post they accompany. This matters because the framework builds on modern generative tools: the underlying literature the authors draw on includes latent diffusion models for high-resolution image synthesis and bootstrapped language-image pre-training approaches, techniques that can produce plausible synthetic imagery but require careful control when the goal is faithful relation extraction rather than creative image generation.

Another key element is the Role-Aware Attention mechanism, which brings textual knowledge to bear on the visual stream. In a social media post, the entities of interest play specific roles: one may be the subject of an action and the other its object. DNMMA explicitly incorporates these textual entity roles when processing the image, allowing the model to suppress visual features that are noisy or irrelevant to the entities in question. If a post mentions two politicians and the attached photo shows them at a podium surrounded by reporters and flags, the attention mechanism can learn to focus on the regions depicting the two individuals and their interaction, downplaying the background elements. Complementing this, a Salient Visual Augmentation component strengthens the discriminative power of local key regions, sharpening the model’s ability to distinguish the visual evidence that actually supports one relationship type over another. Object detection and visual grounding research, including open-set detection systems referenced in the paper’s bibliography, provides the technical backdrop for locating these salient regions.

Once both modalities have been refined, DNMMA performs joint entity relation optimization, integrating the cleaned textual and visual representations to make the final relation prediction. The architecture thus follows a clear pipeline: first stabilize the text, then enrich and filter the image at multiple granularities, then fuse the two streams under the guidance of entity roles, and finally optimize the relation classification jointly across the modalities. This staged design reflects a broader lesson from the multimodal machine learning literature, where researchers have repeatedly found that simply concatenating text and image features performs poorly compared with approaches that respect the distinct noise profiles and informational roles of each modality. The authors’ framework can be read as a systematic application of that lesson to the specific challenges of social media content.

The experimental evaluation was carried out on two benchmark datasets, MNRE and MRE-MI, both of which are publicly available. MNRE, introduced in 2021, is a challenging multimodal dataset for neural relation extraction with visual evidence in social media posts, and it has become the standard testing ground for this line of research. MRE-MI, published more recently, extends the task to posts containing multiple images, a setting that better reflects real-world social media but also amplifies the noise and alignment problems the new framework is designed to solve. According to the paper, extensive experiments on both datasets demonstrate the effectiveness of DNMMA in social media multimodal relation extraction, with the dual-modality noise mitigation and multi-granularity augmentation components each contributing to the overall performance gains.

The significance of this work extends beyond a single benchmark. Relation extraction is a foundational technology for knowledge graph construction, the process of building machine-readable networks of entities and their relationships that power search engines, question-answering systems, and recommendation platforms. As the field’s own survey literature notes, relation extraction has entered a new era in which large language models and multimodal inputs are reshaping what is possible. Yet the web’s most dynamic relational data, the constant chatter of social media, remains difficult to harvest because of its low quality. A framework that explicitly models and mitigates noise in both text and images, while augmenting the visual modality at multiple granularities, offers a template for making that harvest practical. The authors’ emphasis on fine-grained alignment, in particular, addresses a recognized gap: earlier systems that relied on coarse visual features were often injecting as much confusion as evidence.

The publication also situates itself within a rapidly growing research conversation. The bibliography spans work on text-image relation propagation models, hierarchical visual prefixes for extraction, contrastive alignment methods, variational information bottlenecks for denoising, and mixup-based image augmentation for multimodal named entity recognition. DNMMA synthesizes ideas from these threads into a coherent whole, and its dual focus, mitigating noise in both modalities rather than treating the image as the sole culprit, marks a conceptual shift in how the community frames the problem. The authors declare no competing financial interests and report that no funding was received for the preparation of the manuscript, and the datasets underpinning the evaluation are openly available, which should make the results straightforward for other groups to scrutinize and extend.

For the broader public, the work hints at a future in which the flood of mixed text-and-image content online can be automatically organized into structured knowledge rather than simply archived. Better relation extraction means better answers to questions like who attended an event, which companies are linked to which products, or how communities relate to public figures, all derived from the informal, image-laden posts that dominate contemporary communication. It also raises the familiar stakes of automated social media analysis: the same machinery that builds knowledge graphs could be applied to monitoring, profiling, or misinformation detection, making the technical quality of these systems a matter of broad societal interest. With DNMMA, Zhang and Guo have shown that confronting the noise problem on both fronts, the written word and the accompanying picture, is not merely a cleanup exercise but a route to substantially more reliable machine understanding of the social web. The paper was received in February 2026, accepted in September 2026, and published on 1 October 2026 in volume 68 of Knowledge and Information Systems.

Subject of Research: A dual-modality noise mitigation and multi-granularity augmentation framework for multimodal relation extraction in social media posts

Article Title: Dnmma: dual-modality noise mitigation and multi-granularity augmentation for multimodal relation extraction in social media

Article References: Zhang, Y., & Guo, Y. (2026). Dnmma: dual-modality noise mitigation and multi-granularity augmentation for multimodal relation extraction in social media. Knowledge and Information Systems, 68(1), Article 271. https://doi.org/10.1007/s10115-026-02905-z

Image Credits: AI Generated

DOI: 10.1007/s10115-026-02905-z

Keywords: multimodal relation extraction, social media, knowledge graphs, noise mitigation, data augmentation, attention mechanism, text-image alignment, information extraction, machine learning, MNRE dataset, visual augmentation, natural language processing

The imperial workshops of the Qing court, administered through bodies such as the Zaobanchu, generated enormous quantities of routine documentation. Memorials, requisitions, work logs, and delivery records capture the names of craftsmen, the materials they used, the deadlines imposed by the palace, and the judgments of supervising officials. For historians, this material is extraordinarily rich but also extraordinarily difficult to mine. The documents were written in classical Chinese with period-specific terminology, honorific conventions, and abbreviations, and the sheer volume of records makes manual extraction of structured facts a decades-long endeavor for even a well-funded team of archivists.

Knowledge graphs have become a standard tool for organizing such information. A knowledge graph represents entities, for example a craftsman, an object type, a workshop department, or a date, as nodes, and expresses relationships between them as typed edges, such as produced, supervised, requested, or made of. Once built, a graph allows researchers to ask structured questions: which craftsmen worked on both enamel and lacquer in a given year, which materials were most frequently requested for objects destined for particular palace halls, or how production patterns shifted after the death of one supervisor and the appointment of another. The difficulty has always been populating the graph, a task that traditionally required hand-coding rules or annotating large volumes of text by hand.

The researchers turned to large language models to automate this extraction. Modern LLMs, trained on enormous corpora of text, have demonstrated a strong ability to perform information extraction tasks with minimal task-specific training, a property that makes them attractive for low-resource domains such as classical Chinese archival documents where annotated training data is scarce. Rather than training a bespoke model from scratch, the approach described in the study leverages prompting strategies that instruct a general-purpose language model to identify entities and relations in archival passages and to normalize them into a consistent schema designed for the domain of imperial craft production.

Building a reliable pipeline required more than prompting alone. The authors confronted classic problems of LLM-based extraction: hallucinated entities, inconsistent naming of the same person or object across documents, ambiguity in classical Chinese phrasing, and the need to map colloquial or archaic terms onto a controlled vocabulary. Their framework therefore combines the generative strengths of language models with conventional knowledge engineering practices, including schema design, entity normalization and deduplication, and validation steps that check extracted records for internal consistency before they are committed to the graph. The result is a hybrid workflow in which the language model does the heavy lifting of reading unstructured text, while structured constraints keep the output tidy enough for scholarly use.

The choice of the Yongzheng reign as the target domain is historically motivated. Yongzheng, who reigned from 1722 to 1735, is known for close personal supervision of imperial manufacturing and for the high quality and distinctive aesthetic of objects produced under his direction. The workshop records from his reign are comparatively well preserved and internally consistent, which makes them an excellent testbed for computational methods: the density of documented production events provides ample material for extraction, while the known historical context offers a way to sanity-check the resulting graph against what historians already know about the period.

The resulting knowledge graph, once assembled, functions as a navigable map of the imperial production system. Nodes capture not only individual artifacts and their materials but also the institutional geography of the workshops, the networks of craftsmen, and the temporal rhythm of commissions. Such a resource supports quantitative analysis that would be impractical with the raw documents alone. Researchers can trace the lifecycle of an object from initial imperial request through design approval, material allocation, fabrication, revision, and final delivery, and can aggregate thousands of such lifecycles to reveal systemic patterns, for instance seasonal fluctuations in production or the concentration of particular techniques within particular workshop divisions.

Beyond Qing studies, the significance of the work lies in its demonstration of a generalizable method for the digital humanities. Many of the world’s historical archives remain locked in unstructured text, and the combination of large language models with knowledge graph construction offers a template that can be adapted to other corpora, languages, and domains. The study also contributes to an ongoing conversation about rigor in LLM-assisted scholarship. Because generative models can produce plausible but incorrect extractions, the authors emphasize evaluation of extraction quality and the importance of keeping human experts in the loop, particularly for contested or ambiguous readings of historical documents. The graph is best understood not as a replacement for close reading but as an index and hypothesis generator that directs expert attention to the most promising passages.

Limitations acknowledged in work of this kind are worth noting. Extraction accuracy for classical Chinese remains imperfect, entity resolution across hundreds of thousands of records is an open problem, and any schema imposes interpretive choices that shape downstream analysis. The authors position the current graph as a foundation to be extended and refined, with future directions including richer temporal modeling, integration with museum collection databases and surviving physical artifacts, and multimodal linkage between archival text and images of the objects themselves. As large language models continue to improve in multilingual and historical-language capability, pipelines of this type are likely to become a standard part of the historian’s toolkit.

For readers of science news, the study is a vivid example of artificial intelligence moving beyond chatbots and code generation into the archive. A language model trained largely on modern text was adapted to read eighteenth-century palace paperwork and return a structured, queryable model of an entire imperial manufacturing economy. In doing so, it turns stacks of memorials and requisition slips into something closer to a database of the Yongzheng court’s creative life, allowing scholars to see the craftsmen, materials, and decisions behind the treasures that now anchor museum collections worldwide, and pointing toward a future in which the deep past becomes searchable at scale.

Subject of Research: Large language model-driven knowledge graph construction for Qing Yongzheng imperial artifact production

Article Title: Large Language Model-driven knowledge graph construction for Qing Yongzheng imperial artifact production

Article References: Large Language Model-driven knowledge graph construction for Qing Yongzheng imperial artifact production. (n.d.). https://doi.org/10.1038/s41599-026-09040-8

Image Credits: AI Generated

DOI: 10.1038/s41599-026-09040-8

Keywords: large language models, knowledge graph, Qing dynasty, Yongzheng reign, imperial artifact production, digital humanities, information extraction, classical Chinese archives, Zaobanchu, entity normalization, cultural heritage computing, Large

The core idea behind MR-UIE is a shift from treating large language models as passive extractors to training them as active reasoners. Rather than simply asking a model to map input text onto a fixed output structure, the framework teaches it to consider multiple analytical perspectives, weigh different reasoning strategies, and select the path most likely to yield a correct, faithful extraction. The researchers describe this as enabling models to understand not only what to extract but also how to reason, a distinction they argue is critical when texts are riddled with ambiguity, implicit causal links, and context-dependent information that simple pattern matching cannot resolve.

Technically, the framework rests on four pillars. The first is a unified schema representation, drawing on function-call semantics, that converts task-specific schemas for entities, relations, and events into a common hierarchical JSON structure with class identifiers, property-defining arguments, and semantic descriptors. This abstraction allows a single model to handle named entity recognition, relation extraction, and event extraction within one consistent output grammar, and it lets schemas for unseen tasks be projected into the same space at inference time, supporting zero-shot generalization without task-specific adapters.

The second pillar is the construction of a multi-perspective reasoning dataset, modeled on a human cognitive cycle of brainstorming, generating multiple solutions, and converging on the best one. For each input text and schema, a large language model acting as a strategy generator produces fine-grained thinking strategies along three dimensions: cognitive perspective, professional role, and heuristic rules. These raw strategies are then clustered into keyword-defined paradigms, ranked by TF-IDF embeddings and cosine similarity, and sampled to balance distinctiveness against representativeness. DeepSeek-R1 is then prompted to produce step-by-step chain-of-thought rationales together with structured predictions for each retained strategy, and rejection sampling discards any strategy whose prediction disagrees with the ground truth. Samples that fail to accumulate a minimum number of correct strategies are withheld from supervised training and routed instead to the reinforcement learning stage, ensuring the model learns from cases where reasoning demonstrably helps.

Training proceeds in three stages. The model, built on Qwen3-8B, first undergoes supervised fine-tuning on a curated corpus of roughly one million instances drawn from the bilingual IEPile collection, teaching it to interpret instructions and emit outputs that strictly conform to the unified schema. In the second stage, multi-perspective reasoning fine-tuning injects strategy prefixes into inputs and introduces special thinking tokens that segment reasoning steps, with a composite objective that optimizes both the chain-of-thought rationales and the structured predictions. To prevent overfitting to any single strategy, the team employs a regularization trick called strategy hiding, in which ten percent of training examples instruct the model to skip reasoning entirely. The final stage applies reinforcement learning using a rule-based Group Relative Policy Optimization procedure, in which the model generates multiple candidate reasoning chains per question and learns from a multi-grained reward function that scores both results and process.

The reward design is notable for its granularity. The result reward decomposes correctness into category-level and argument-level components, reweighted to reflect the fact that category prediction is easier than argument extraction, and combined as a weighted harmonic mean to encourage joint optimization. The process reward evaluates faithfulness, checking whether the reasoning strictly follows the schema, effectively uses the source text, and reflects a sound strategy. A convex combination of the two guides the policy, allowing the model to internalize effective reasoning patterns rather than memorize specific templates. During reinforcement learning, difficult samples are deliberately emphasized to raise the ceiling of model capability while leaving reasoning formats unconstrained, encouraging autonomous exploration of optimal strategies.

The experimental results are striking. Across eleven supervised benchmarks spanning newswire, finance, biomedicine, and general science, MR-UIE set new state-of-the-art Micro-F1 scores on five datasets, and on the remaining six never trailed the best published result by more than 3.2 F1 points. The wins span the full spectrum of extraction tasks: the framework topped the mit-restaurant dataset with its eight colloquial entity types and ambiguous boundaries, led on the terminology-dense biomedical AnatEM corpus despite limited training data, achieved the best overall scores on the skewed eighteen-category KBP37 relation benchmark, and ranked first on argument extraction for DUEE-fin, a financial event dataset with thirteen event types, ninety-one argument roles, and severe role overlap.

Zero-shot performance tells perhaps the more important story. Evaluated on datasets from domains and schemas never seen during training, spanning literature, music, law, and social media in both English and Chinese, MR-UIE achieved top-two results on most tasks and beat GPT-4 in five experimental settings, including a new state-of-the-art for zero-shot named entity recognition on CrossNER. The researchers attribute this generalization to the unified schema, which lets supervised capabilities transfer without structural adaptation, and to adaptive chain-of-thought reasoning that dynamically decomposes problems into finer-grained sub-steps, helping the model distinguish easily confused elements on schema-heavy datasets.

Ablation studies confirm that each component earns its place. Removing the unified schema cost between 0.42 and 0.72 F1 points across subtasks, while dropping multi-perspective reasoning training caused a maximum decline of 2.48 F1 points, demonstrating that explicit multi-angle fine-tuning is essential rather than decorative. Reinforcement learning alignment proved to internalize the diverse strategies rather than memorize them: when compared directly against static alternatives, the reinforcement-learned dynamic strategy beat random template selection, which actually degraded performance below a no-strategy baseline, and outperformed similarity-based template matching, whose effectiveness varied by task. Intriguingly, the training dynamics showed that as reinforcement learning progressed, generated reasoning chains shortened and converged to roughly 300 tokens while reward scores steadily climbed, suggesting that concise, efficient thinking benefits extraction performance more than lengthy elaboration.

The authors are candid about limitations, noting that the framework still struggles to dynamically integrate diverse reasoning paths and to handle highly noisy text, and they point toward noise-aware reasoning modules, contrastive path selection, refined reward design, and extension to multimodal settings as future directions. Even so, the work makes a compelling case that the road to robust structured understanding runs through reasoning itself. By rewarding models for faithful, efficient thinking and giving them multiple analytical lenses to think through, MR-UIE demonstrates that teaching language models how to reason may matter as much as teaching them what to know, a conclusion likely to shape how the next generation of extraction systems is built.

Subject of Research: Multi-perspective reasoning with reinforcement learning for universal information extraction

Article Title: MR-UIE: multi-perspective reasoning with reinforcement learning for universal information extraction

Article References: Li, Z., Wang, S., Fang, R., Bao, M., Wu, Z., Song, S., Li, Y., & He, Z. (2025). MR-UIE: multi-perspective reasoning with reinforcement learning for universal information extraction. Vicinagearth, 2(1), Article 17. https://doi.org/10.1007/s44336-025-00019-8

Image Credits: AI Generated

DOI: 10.1007/s44336-025-00019-8

Keywords: information extraction, multi-perspective reasoning, reinforcement learning, large language models, chain-of-thought, universal information extraction, named entity recognition, relation extraction, event extraction, GRPO, zero-shot generalization, schema representation