Trustworthy Domain-Specific AI for Structured Knowledge Retrieval and Reasoning

Ryan C. Barron

cs.IROct 6, 2026

Trustworthy Domain-Specific AI for Structured Knowledge Retrieval and Reasoning

Authors: Ryan C. Barron

Abstract

This dissertation presents a scalable architecture for transforming unstructured, domain-specific text into structured knowledge for retrieval and reasoning. It integrates semi-automatic corpus curation, semantic structuring, retrieval, and inference into an interpretable pipeline. The research introduces Binary Bleed, an adapted binary search method that reduces low-rank search complexity for Non-negative Matrix Factorization (NMF), and Hierarchical NMF with automatic latent feature selection (HNMFk), a depth-adaptive topic modeling method that produces interpretable taxonomies guided by subject matter experts. These representations populate a typed Knowledge Graph and a semantically aligned Vector Store containing extracted latent features, synchronized through an event-driven substrate. Tensor-Structured Retrieval-Augmented Generation (T-SRAG) dynamically routes queries across retrieval paths. Contrastive alignment maps document and query embeddings to hierarchical topic structures to improve semantic fidelity and reduce hallucinations. Beyond retrieval, tensor-based link prediction identifies and completes missing links in the Knowledge Graph, supporting inference grounded in citation structure. Applications across cybersecurity, law, materials science, and healthcare demonstrate improvements in retrieval precision, early trend detection, hypothesis generation, and hallucination mitigation. The dissertation provides a deployable, modular foundation for trustworthy, domain-specific AI systems that retrieve and reason over structured knowledge.

Figures & tables

Acronym	Definition
AI	Artificial Intelligence
ANN	Approximate Nearest Neighbour
BERT	Bidirectional Encoder Representations from Transformers
BUNIE	Bibliographic Utility Network Information Expansion
BM25	Best Matching 25
BMF	Boolean matrix factorization

Table 1 : Acronyms used in this dissertation

Figure 1.1 : End-to-end pipeline: SME-seeded corpus $\rightarrow$ Binary Bleed rank search $\rightarrow$ HNMFk topic tree $\rightarrow$ synchronised KG + vector store $\rightarrow$ hybrid RAG agent.

Figure 4.1 : BUNIE’s distillation pipeline. Panel A : The SME selects specific papers, shown as a bag-of-words wordcloud. Panel B : The dataset expands through the citation network. The wordcloud of a selected subset shows how the vocabulary differs from the core. Panel C : Human-in-the-loop pruning removes data. The wordcloud begins to resemble the core. Panel D : An embedding heuristic removes papers far from the core. The dataset becomes more compact. The wordcloud approaches parity with the original. Panel E : NMFk topic modeling produces H-clustering factorization to remove clusters lacking a core paper. A dense, yet substantial set of documents indicates a successful distillation. The wordcloud closely resembles the original. As indicated by the Ellipse , iterations can be made for refinement or the data extracted in repose for downstream analysis.

Figure 4.2 : 2-dimensional representation of hypersphere pruning. Core papers are red dots, citations/ references are blue squares, and hyperspheres are dotted red circles.

Step	Line	Complexity
Initialization	1–2	$O(1)$
Base Case	3–4	$O(1)$
Middle/Check	5–7	$O(1)$
Model/Score	8	$T_{\text{M}}+T_{\text{S}}$
Update Min/Max	10–16	$O(1)$
Bound Check	17,20	$O(1)$

Figure 5.1 : Primitive operations for asymptotic analysis.

Panel of Figure 5 — Figure 5.2 : Traversal sorts: preorder (red), inorder (green), postorder (blue).

Order	Operation 1	Operation 2
T1	Traversal Order Sort	Chunk Ks by Resource Count
In	1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11	[1, 2, 3, 4, 5, 6] [7, 8, 9, 10, 11]
Pre	6, 3, 2, 1, 5, 4, 9, 8, 7, 11, 10	[6, 3, 2, 1, 5, 4] [9, 8, 7, 11, 10]
Post	1, 2, 4, 5, 3, 7, 8, 10, 11, 9, 6	[1, 2, 4, 5, 3, 7] [8, 10, 11, 9, 6]
T2	Traversal Order Sort	Chunk Ks by Alg. 4
In	1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11	[1, 3, 5, 7, 9, 11] [2, 4, 6, 8, 10]

Table 5.1 : Chunks Pre/Post Traversal Sorts on two resources

Node labels
Document	{doi,title,year,version}
Author	{author_id,name}
Affiliation	{aff_id,name,country}
Venue	{venue_id,name}
Topic	{topic_id,level,coherence}
Keyword	{term}

Table 6.1: Core KG schema with provenance and monotone version properties.

Notation	Description	Notation	Description
$x$	Scalar	$\mathbf{x}_{i}$	$i$ th element in the vector
$\mathbf{x}$	Vector	$\mathbf{X}_{ij}$	Entry at row $i$ , column $j$
$\mathbf{X}$	Matrix	$\mathbf{X}_{i:}$	$i$ th row
$\boldsymbol{\mathscr{X}}$	Tensor	$\mathbf{X}_{:j}$	$j$ th column
$\boldsymbol{\mathscr{X}}_{::i}$	$i$ th slice (3rd dim)	$\boldsymbol{\mathscr{X}}_{:::i}$	$i$ th slice (4th dim)
$\boldsymbol{\mathscr{X}}^{\textit{name}}$	Superscript identifier	$*$	Dot product

Table 7.1: Summary of the notation styles used in the paper.

Panel of Figure 8 — Figure 8.1 : NMFk (top row) and K-means (bottom row), Vanilla (Left) and Early stopping (Right). $\forall k_{optimal}=k_{true}$ .

Figure 8.3 : Binary Bleed reduction on distributed Rescal (Purple), and distributed NMF (Green).

Figure 8.4 : Word cloud of selected topics and their interpreted labels. The method is hierarchically applied three times, selecting a topic to extract its subtopics at each level. Selected topics from each level are shown in each row of the figure.

Figure 8.5 : Distribution of the top ten document categories, based on the arXiv author assignments. Each topic corresponds to the word clouds from Figure 8.4 . Selected topics from each level are shown in each row of the figure.

Figure 8.6 : New Mexico Supreme/Appeals case counts per year.

Figure 8.7 : 10 latent topics from Constitutional Provisions .

Figure 8.8 : 985 latent topics from Statutory Sections .

Figure 8.9 : 420 latent topics from Court of Appeals Cases .

Figure 8.10 : 248 latent leaf topics from Supreme Court Cases .

#	Constitution Label
0	Irrigation and Water Resource Management Principles
1	Regulation of Private Sector Influence on Public Schools and Education Services
2	Branches of Government Structure and Functionality
3	Education Funding for New Mexico’s Educational Institutions
4	Legislative Proceedings and Lawmaking Activities Enacted During Sessions
5	Territorial Transition: Constitutional Ratification and Statehood Provisions

Table 8.1 : NM Constitutional Depth-0 H-Clusters

#	Statute Label
0	Municipal Court Civil Cases Involving Children’s Rights
1	Public Education Infrastructure Management Systems
2	Criminal Codes, Local Governance Boundaries, Licensing Rules
3	Comprehensive Emergency Health and Human Services Response Framework
4	Taxation and Revenue Collection Oversight
5	Military Decorations, Licensing Procedures, Governance Boards

Table 8.2 : NM Statutory Depth-0 H-Clusters

#	Appeals Court Label
0	Real Property Rights and Interests
1	Civil Liability and Injury Issues Arising from Healthcare Services
2	Parental Rights and Custody Proceedings Involving Disputed Parental Fitness
3	Motor Vehicle Insurance and Liability Claims Processing Procedures
4	Mortgage Foreclosure and Secured Lending Frameworks
5	Arraignments, Tribal Jurisdiction, Divorce, Bond Conditions, Motor Licensing

Table 8.3 : NM Appeals Court Case Law Depth-0 H-Clusters

#	Supreme Court Label
0	Arbitration of Contract Disputes and Judicial Decision-Making in Motor Vehicle Cases
1	Revenue and Taxation Frameworks in Governance and Administration
2	Mineral Rights Leases
3	Native American Self-Governance and Tribal Jurisdictional Frameworks
4	Municipal Zoning Ordinances and Regulations of Local Governance Areas
5	Damages Award for Wrongful Conduct Against Business Partners

Table 8.4 : NM Supreme Court Case Law Depth-0 H-Clusters

Node Type	Nodes	Out Edges	Legal Cites
NMFk Topics	2,469	92,634	–
NMFk Keyword	11,076	8,281,843	–
BOW Vocabulary	132,423	–	–
Constitution	265	9,067	41
Statute	28,251	1,930,707	81,353
Supreme Court Case	5,727	2,437,161	76,478

Table 8.5 : Neo4j Node and Edge Overview

Figure 8.11 : Examination of ‘Estoppel’ relating to being a keyword in topics, vs bag of words vocabulary.

Figure 8.12 : 25 questions queried across five different LLM channels. “Attempts” indicate the percentage of responses that tried to answer the question, and “accuracy” represents the percentage of correct attempts.

Figure 8.13 : Comparison of metrics (ROUGEL, NLI entailment, SummaC, and FactCC) across models: Nemotron–70B-Instruct, Claude 3 Opus, GPT-4o, Gemini Pro, and our system.

#	Label	# Docs.	Percent
0	Malware Behavioral Analysis	158	1.80
1	Cybersecurity Challenges	305	3.47
2	Cybersecurity Research	114	1.30
3	Botnet Detection Techniques	142	1.62
4	Malware Feature Selection And Extraction	353	4.02
5	Network Intrusion Detection	134	1.52

Table 8.6 : Labels for Topic Clusters

Figure 8.14 : The KG schema. Images generated with DALL·E [ 170 ] .

Figure 8.15 : Keyword ’cybercrime’ graph search. A single keyword (green) and linked documents (light blue) are returned. The documents also link affiliated institutions (yellow) and the country of the institutions (red).

Properties/Dataset	H.sapiens-extended	Brain	Disease of Metabolism	Liver	Neurodegenerative Disease
Number of proteins	14,407	11,167	1,036	10,627	820
Number of positive pairs	157,950	225,200	5,131	218,239	5,881
Number of negative pairs	157,300	223,130	5,123	215,984	5,879

Table 9.1 : The statistics of the five PPI datasets include the number of proteins and positive and negative interaction pairs of each PPI network when represented as a matrix.

H.sapiens - extended
Dataset / Metric	BNMFk	RNMFk	WNMFk	$\text{BNMFk}_{\text{lmf}}$	$\text{RNMFk}_{\text{lmf}}$	$\text{WNMFk}_{\text{lmf}}$	LMF	symLMF
ROC AUC	0.755 ± 0.022	0.941 ± 0.001	0.951 ± 0.003	0.959 ± 0.002	0.955 ± 0.001	0.955 ± 0.001	0.949 ± 0.001	0.944 ± 0.001
PR AUC	0.884 ± 0.009	0.957 ± 0.001	0.964 ± 0.003	0.969 ± 0.002	0.965 ± 0.001	0.965 ± 0.001	0.934 ± 0.001	0.955 ± 0.001
UQ – ROC AUC	0.762 ± 0.027	0.942 ± 0.001	0.954 ± 0.003	0.962 ± 0.004	0.956 ± 0.001	0.957 ± 0.001	–	–
UQ – PR AUC	0.880 ± 0.010	0.958 ± 0.001	0.966 ± 0.002	0.969 ± 0.004	0.966 ± 0.001	0.967 ± 0.001	–	–
Brain

Table 9.2 : Performance comparison of BNMFk, RNMFk, WNMFk, and their LMF - based extensions across five protein–protein interaction datasets. Bold values indicate the best classical score per block; dashes indicate metrics not reported by uncertainty - quantified variants.

Figure 9.1 : Dog Dataset - Four binary images are used as Boolean latent features to generate the synthetic data of shape $400\times 16$ .

Figure 9.2 : Results for Dog Data across methods including

\text{BNMFk}_{\text{{kmeans}}}

, NMFk, and WNMFk. Boolean thresholding is not used for NMFk and WNMFk. The first row presents violin plots visualizing the rank

k

predictions at different test-set size levels. The second row displays RMSE scores for the test set, demonstrating the missing link prediction performance. The results are reported for each rank

k

on the x-axis, with the dark/dashed vertical line across columns being the true rank

k=4

Figure 9.3 : Results for Dog Data across methods, including BNMFk, NMFk, and WNMFk, evaluated under different Boolean thresholding techniques. The Boolean thresholding techniques are denoted with the subscripts of kmeans , otsu , and search (coordinate descent). The first row presents violin plots visualizing the rank

k

predictions at different test-set sizes. The second row displays RMSE scores for the test set, demonstrating the missing link prediction performance. The results are reported for each rank

k

on the x-axis, with the dark and dashed vertical line across each column representing the true rank

k=4

Figure 9.4 : Results for Dog Data across methods, including BNMFk, NMFk, and WNMFk, evaluated under different Boolean thresholding techniques. The Boolean thresholding techniques are denoted with the subscripts of kmeans , otsu , and search (coordinate descent). The first row presents the change in RMSE after making predictions on the non-abstained samples, where a negative change indicates improved performance (lower RMSE) for the given fraction of abstained samples. The second row illustrates the fraction of abstained predictions, representing the proportion of cases where the model opted not to make a prediction due to uncertainty. The results are reported for each rank

k

on the x-axis, with the dark and dashed vertical line across each column indicating the true rank

k=4

Figure 9.5 : Computation times for each method are shown in the log scale on the Dog Data. The results indicate that methods leveraging coordinate descent (search) for Boolean thresholding require significantly longer computation times than other approaches.

Figure 9.6 : Swimmer Dataset - Dataset of 16 swimmer images. The first and third rows are the images with real-valued intensities ranging from 0 to 19. The second and fourth rows display the Boolean versions obtained after applying Otsu thresholding. For analysis, the Boolean versions of the dataset are represented as a matrix of size $1024\times 256$ .

Figure 9.7 : The performance of each method on the Swimmer Data is shown. The first column features a violin plot displaying the distribution of

k

predictions for each method, with the y-axis representing the methods and the x-axis representing the predicted

k

values. The second through final columns present results at each

k

decomposition value, with the x-axis corresponding to the rank

k

. The dark dashed vertical lines indicate this dataset’s true rank of

k=16

. The Boolean thresholding techniques are denoted with the subscripts of kmeans , otsu , and search (coordinate descent), while the uniform subscript refers to running BNMFk without Boolean thresholding on the latent factors. NMFk and WNMFk, without subscripts, are the results of not using Boolean techniques with these methods.

Figure 9.8 : Results for

k

predictions on Gaussian Data are shown for WNMFk and RNMFk methods across different data sparsity levels, represented by the increasing test-set size on the x-axis. The y-axis indicates the predicted

k

values. The results demonstrate that

k

predictions remain accurate for both methods until high levels of sparsity are reached, while WNMFk results in more accurate

k

predictions. Line color and style differentiate between different true

k

values. For instance, a solid line with circle markers represents decompositions where the matrix rank was

k=2

Figure 9.9 : Link prediction performance on test-set sizes of 10% (red lines) and 50% (blue lines) is shown for WNMFk and RNMFk, with RMSE scores plotted on a log scale. The x-axis represents the

k

values, displaying results at different

k

decompositions. Line markers are used to differentiate between the true

k

values. For example, solid red lines with circle markers represent matrices with a true rank of

k=2

and a test-set size of 10%.

Figure 9.10 : Pearson Correlation Coefficients with CI between UQ and errors at the test set are reported across different sparsity levels or test-set sizes that include 0.1, 0.5, and 0.9 (columns) and true matrix ranks

k

including

k\in[2,3]

(rows). The shared x-axis differentiates between the methods WNMFk and RNMFk using hues, while the shared y-axis represents the Pearson correlation values. Bars with different colors correspond to results at various

k

decompositions.

Figure 9.11 : Relative-Std-to-Mean Ratio of Uncertainty plot provides a quantitative perspective on the behavior of uncertainty estimates across varying sparsity levels (or test set size) for each

k

true. The top figure shows the standard deviation (std) at test points divided by the overall average of certainty (UQ). The bottom figure investigates certainty saturation by normalizing the std-to-mean uncertainty ratio at the top plot with the correlation between the UQ and errors at test points.

Figure 9.12 : Mean hit@1 and hit@3 scores (bars, annotated) of the BNMF

+

LMF ensemble after masking every known superconducting link and an equal number of random zeros for each of the four benchmark compounds

Figure 9.13 : Split-violin plot of posterior probabilities for 24 superconducting edges (green). [See also Table 9.3 .] 192 non-superconducting edges (blue) were added for comparison. (Note that these are identical for the four violins.) Lines denote the median and 25th/75th percentiles, with numbers showing cumulative counts. Superconductors concentrate near 1, non-superconductors near 0, indicating a clear separation.

Aggregated Prediction Scores for Masked Materials
Material	Set	Predicted Score
S 2 Ta	Superconductor	0.813
NbSe 2	Superconductor	0.757
MoS 2	Superconductor	0.703
Se 2 Ta	Superconductor	0.699
FeS 2	Non-superconductor	0.206

Table 9.3: Predicted scores assigned to materials whose labels were fully masked during training. Known superconductors (top block) score significantly higher than non-superconductors (bottom block), with a clear separation near 0.5.

Figure 10.1 : Two question types, document and topic, showing LLM percent attempts and correct with SMART-SLIC RAG and without RAG.

Figure 10.2 : MRR and top 10 hit rate percentage per data category over several embedding methods.

Statistic	Value
Documents	46 862
Hierarchical depth	3
TMD compounds	73

Table 10.1 : TMD corpus statistics.

Explore similar work

May 20, 2026cs.AI

SciAtlas: A Large-Scale Knowledge Graph for Automated Scientific Research

The exponential growth of global academic output has confronted researchers and AI agents with an unprecedented

information explosion,'' where fragmented and unstructured knowledge organization impedes deep interdisciplinary integration. Current academic retrieval tools predominantly rely on superficial keyword matching or vector-space semantic retrieval, which lack the topological reasoning capabilities required to navigate complex logical connections. Agentic deep-research-based frameworks are often prone to logical hallucinations and consuming high inference costs. To bridge this gap, in this report, we introduce SciAtlas, a large-scale, multi-disciplinary, heterogeneous academic resource knowledge graph designed as a panoramic scientific evolution network. By integrating over 43M papers from 26 disciplines, and a total of 157M entities and 3B triplets, SciAtlas provides a structured topological cognitive substrate that dismantles disciplinary barriers and furnishes AI agents with a global perspective. Furthermore, we develop a neuro-symbolic retrieval algorithm featuring tri-path collaborative recall and graph reranking, achieving a seamless transition from simple semantic matching to deterministic association discovery. We also present key application directions of SciAtlas, including literature review, automated research trend synthesis, idea positioning, and academic trajectory exploration, to demonstrate that SciAtlas can serve as an effective

cognitive map'' to empower the full loop of automated scientific research while significantly reducing reasoning costs. We have released the interfaces for KG retrieval and various downstream tasks in our GitHub repo.

Shuofei Qiao, Yunxiang Wei, Jiazheng Fan +8

Zhejiang University · University College London

Scholar Information Database Knowledge Graphs Scientific Discovery Neuro-Symbolic Framework Topology

Ask PDF

Apr 24, 2026cs.CL

STEM: Structure-Tracing Evidence Mining for Knowledge Graphs-Driven Retrieval-Augmented Generation

Knowledge Graph-based Question Answering (KGQA) plays a pivotal role in complex reasoning tasks but remains constrained by two persistent challenges: the structural heterogeneity of Knowledge Graphs(KGs) often leads to semantic mismatch during retrieval, while existing reasoning path retrieval methods lack a global structural perspective. To address these issues, we propose Structure-Tracing Evidence Mining (STEM), a novel framework that reframes multi-hop reasoning as a schema-guided graph search task. First, we design a Semantic-to-Structural Projection pipeline that leverages KG structural priors to decompose queries into atomic relational assertions and construct an adaptive query schema graph. Subsequently, we execute globally-aware node anchoring and subgraph retrieval to obtain the final evidence reasoning graph from KG. To more effectively integrate global structural information during the graph construction process, we design a Triple-Dependent GNN (Triple-GNN) to generate a Global Guidance Subgraph (Guidance Graph) that guides the construction. STEM significantly improves both the accuracy and evidence completeness of multi-hop reasoning graph retrieval, and achieves State-of-the-Art performance on multiple multi-hop benchmarks.

Peng Yu, En Xu, Bin Chen +2

AI Product Center, Kingsoft Corporation, Beijing, China

Graph-Based Retrieval Knowledge Graphs Multi-Hop Reasoning

Ask PDF

May 22, 2026cs.LG

SeedER: Seed-Expand-Retrieve for Efficient Knowledge Graph Retrieval

Knowledge graphs (KGs) offer a rich representation for relational knowledge, but their irregular structure makes retrieval challenging: ego-graph expansion grows rapidly, and dense embedding methods struggle with multi-hop compositional queries. Several approaches use LLM agents to explore the KG, analyze candidate nodes, and decide where to explore next. While expressive, these approaches can incur substantial computational and memory costs. On the other hand, we show theoretically that dense embeddings precomputed for graph nodes, even with augmented structure and neighborhood-aware features, can require embedding dimensions comparable to the size of the graph to answer families of knowledge graph queries. This limitation can be overcome with query-adaptive embeddings under certain conditions, and there are graph neural network (GNN) variants that can do so. However, processing the whole graph with a GNN can also incur substantial memory and computational costs, and it requires dense ground-truth labels indicating whether each node answers the query. Straightforward

k

-hop selection around anchor nodes can also be problematic: small

k

limits the nodes we can see, and even with small

k

values such as three and four, the

k

-hop subgraph can grow substantially. In this work, we devise a new approach using Graph Transformers and reinforcement learning that is much less demanding in memory and computation, can run on a CPU at inference time as a first-stage retriever, and requires feedback on only a small number of nodes at a time during training. We show that this method is competitive with methods that fine-tune LLMs to retrieve information from KGs. We call our method SeedER (Seed-Expand-Retrieve), and position it primarily as a first-stage retriever that can run with modest computational resources.

Hamed Shirzad, Frederik Wenkel, Dominique Beaini +2

Valence Labs, Montréal, QC, Canada · Recursion, Salt Lake City, UT, USA · University of British Columbia, Department of Computer Science, Vancouver, BC, Canada

Graph-Based Retrieval Knowledge Graphs Multi-Hop Reasoning

Ask PDF