GlyRAG: Context-Aware Retrieval-Augmented Framework for Blood Glucose Forecasting

Hassan Ghasemzadeh; Shovito Barua Soumma

REVIEW 5 major objections 5 minor 42 references

GlyRAG: Context-Aware Retrieval-Augmented Framework for Blood Glucose Forecasting

T0 review · 5 major / 5 minor · reviewed 2026-08-03 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read GlyRAG claims that language-model-generated summaries of CGM traces, fused with retrieved similar episodes, reduce blood glucose forecast error at 30–60 minutes without extra sensors.

desk verdict The idea—LLM-generated morphology summaries plus retrieval for CGM forecasting—is worth a look, but the headline numbers don't match the tables, and the 'CGM-only' context actually includes insulin and carb data. read the letter →

arxiv 2601.05353 v2 pith:DJ6G6VQF submitted 2026-01-08 cs.LG cs.ITmath.IT

Shovito Barua Soumma , Hassan Ghasemzadeh This is my paper

classification cs.LGcs.ITmath.IT

keywords bloodglucoseforecastingCGMretrieval-augmentedgenerationlargelanguagemodelstime-seriesclinicalerrorgriddiabetesmanagementmorphologysummarization

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

GlyRAG is claimed to improve long-horizon blood glucose forecasts by having a large language model turn a three-hour CGM window into a short text description of its shape—rising, falling, oscillating, hypo/hyperglycemia risk—embedding that text alongside the glucose values, and then retrieving similar past episodes for cross-attention. On two real-world type 1 diabetes datasets, the paper reports lower RMSE at 5, 30, and 60 minutes than a glucose-only baseline with the same backbone; the abstract's headline numbers on one cohort drop from 13.8 to 10.6 at 30 minutes and 23.1 to 20.2 at 60 minutes, while the body tables show smaller but consistent gains. The paper also claims most predictions land in clinically safe error-grid zones (about 85% in zones A and B) and hypoglycemia sensitivity improves, all without needing extra sensors such as meal or exercise logs. A sympathetic reader would care because usable 30–60-minute forecasts are the clinically valuable range, and a purely CGM-driven way to obtain them would be simpler to deploy than multi-sensor systems.

What carries the argument

Three components carry the argument. (1) An LLM contextualization agent that takes a prompt containing the 3-hour CGM window plus a 30-minute summary of carbohydrate intake, insulin boluses, current BGL, time-in-range and trend, and returns a five-sentence qualitative summary. (2) A multimodal transformer encoder that embeds that summary with a pre-trained language model, patch-embeds the glucose sequence, and fuses the two tokens with multi-head self-attention, trained with a cross-translational loss that aligns the two embeddings. (3) A retrieval-augmented adapter that indexes fused embeddings of training samples, retrieves the top-3 nearest neighbors by cosine similarity, and applies per-

What would settle it

Run the identical GlyRAG pipeline with (a) the LLM prompt stripped of all non-CGM fields, and (b) a template-based summary instead of the LLM summary, on the same two cohorts. If the RMSE improvement over the BGL-only baseline collapses in either condition, the central claim—that CGM morphology summaries and case retrieval create the gain—is refuted. A second check: compare patient-wise RMSE gains to see if they are concentrated in a few subjects.

Watch

Extended reading notes

Core claim

The central claim is that a context-aware, retrieval-augmented architecture—LLM-generated morphological summaries fused with patched glucose embeddings, plus a retrieval module that attends to the top-K similar training episodes—outperforms both a glucose-only baseline and prior published models on two type 1 diabetes benchmarks, especially at 30- and 60-minute horizons, and that the improvement is clinically meaningful: around 85% of predictions fall in the two safest zones of a standard clinical error-grid analysis, and hypoglycemia sensitivity exceeds the strongest prior model on one cohort. The claim is explicitly CGM-only, needing no additional sensing modalities. The body presents a mo

Load-bearing premise

The entire 'CGM-only' story rests on the belief that the LLM's summaries of glucose morphology—not the carbohydrate, insulin, and time-in-range numbers that the prompt also supplies—are what drives the forecast improvement. If that belief is wrong, the comparison against BGL-only baselines is unfair.

Editorial extensions

If this is right

Long-horizon (30–60 min) glucose forecasts can be improved without new sensors, if the claim holds, making decision-support tools easier to deploy.
The effect appears to come from general LLM reasoning ability, since two different language models both produced gains.
Retrieval of similar past episodes is a load-bearing component: ablations show the largest drop when it is removed.
Hypoglycemia sensitivity and error-grid safety improve, not only raw RMSE, which matters for clinical use.
The pipeline provides interpretable natural-language summaries that could be shown to patients or clinicians.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

The headline improvement may be partly driven by non-CGM fields in the prompt (carb intake, insulin bolus, time-in-range); if so, the 'CGM-only' framing overstates the result. A cleaner test would strip those fields from the prompt and measure the drop.
A hand-written rule-based summary with the same qualitative phrases could serve as an ablated control; if it performs as well as the LLM summary, the gain is from morphology content, not language-model priors.
The retrieval module could be ported to other physiological forecasting problems (say, heart-rate or activity episodes) where similar-past-window matching is natural, but the authors do not explore this.
Patient-level variance in the reported RMSE suggests checking whether the improvement concentrates in a few individuals before treating it as a general population effect.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

The short version: GlyRAG is a plausible new combination—using an LLM to write a short morphology summary of a CGM window, embedding it, fusing it with patch embeddings, and retrieving similar historical episodes to attend over—but the paper overclaims what it shows. The abstract reports 13.8→10.6 and 23.1→20.2 over PatchTST on OhioT1DM; those numbers don't appear anywhere in Table I, and PatchTST isn't evaluated. The actual gains over their own BGL-only baseline are 1–5% RMSE, modest but consistent across horizons.

The combination is genuinely new for CGM forecasting: I don't know prior work doing LLM-based context summarization with retrieval for glucose. The authors compare against several baselines, and the clinical evaluation with Clarke error grid and TIR is a plus. The architecture is clearly described, and the ablation in Table IV suggests both context and retrieval contribute, though the effects are small.

The load-bearing problem is the prompt in Section III-C. It explicitly feeds the LLM carbohydrate intake and insulin bolus fields, and instructs it to analyze glucose 'along with associated carbohydrate intake and insulin administration data.' So the context summaries are not CGM-only, and the baseline (BGL-only) doesn't see any of that. That means the 1–5% gain could come from meal and insulin information rather than morphology. There's no ablation that strips those fields. The abstract's 'without requiring additional sensing modalities' is technically true—carbs and insulin aren't sensors—but 'directly from CGM traces' is not. The stress-test note is right.

Smaller issues: patch length is 4 in Section III-D but 6 in Section IV-E; there are no significance tests or error bars; no code; and the checkmarks in Table IV are hard to parse. The authors list limitations but don't mention this input leak.

Who's this for? Researchers in glucose forecasting or LLM-for-time-series. It's a useful idea to discuss, but the evidence as presented doesn't support the headline claim. With a corrected evaluation, it could be a solid contribution.

My recommendation: send it to peer review. The idea deserves referee time, and the authors need to fix the evaluation. This isn't a desk reject—the task is well-motivated, the approach is coherent, and the flaws are fixable. Just don't trust the abstract numbers.

Referee Report

5 major / 5 minor

Summary. The paper proposes GlyRAG, a retrieval-augmented forecasting framework for blood glucose. A 3-hour CGM window is summarized by an LLM (e.g., GPT-4) into a short textual description; the text is embedded via BERT and fused with patch-embedded CGM values in a multimodal transformer, aided by a cross-translational loss. A retrieval module then finds similar historical training episodes in the fused embedding space and uses cross-attention to produce the final forecast for 5-, 30-, and 60-minute horizons. Evaluation is reported on OhioT1DM and AZT1D, with RMSE/MAE, Clarke Error Grid, CG-EGA, event sensitivity, and TIR comparisons against several baselines. The central claim is that GlyRAG obtains better long-horizon glucose forecasts using context derived purely from CGM traces, without additional sensors or non-CGM inputs, and that the gains are statistically significant.

Significance. If the claims were validated, the idea of using an LLM to generate morphology-aware text summaries from CGM traces and then retrieving analogous episodes would be a useful and interesting contribution to glucose forecasting. The framework is well motivated and the qualitative examples suggest that textual context can capture clinically meaningful patterns. However, the current manuscript does not establish the central claims. The LLM prompt in Section III-C explicitly includes carbohydrate and insulin information, contradicting the 'CGM-only' premise; the abstract quotes numbers that do not appear in Table I; no significance tests or error bars are reported despite the word 'significantly'; and the architectural configuration is described inconsistently across sections. These issues are load-bearing because they directly affect whether the reported improvements are attributable to the proposed CGM-morphology context or to leakage of non-CGM inputs.

major comments (5)

[§III-C, prompt template] The 'CGM-only context' claim is contradicted by the prompt itself. The prompt supplies <Carbohydrate Intake>, <Total Insulin Bolus>, <Food Bolus>, <Correction Bolus>, <Other Bolus>, <Current BGL>, <Time In Range>, <Trend> and instructs the LLM to analyze glucose readings 'along with associated carbohydrate intake and insulin administration data.' The generated text summary is embedded by BERT, fused with the CGM patch embeddings, and used to build the retrieval index. The BGL-only baseline receives none of this non-CGM information. Therefore, the reported gains (e.g., Table I 60-min 20.49→20.22; Table IV 21.59→20.22) may be due to meal and insulin leakage rather than to CGM morphology. No ablation removes these fields from the prompt, so the central 'CGM-derived context' contribution is untested.
[Abstract vs. Table I] The abstract's headline numbers are not supported by the results tables. It claims 'RMSE decreases from 13.8 to 10.6 at 30 minutes and from 23.1 to 20.2 at 60 minutes on OhioT1DM' and claims improvement 'over PatchTST', but Table I contains no PatchTST row, no 13.8 baseline at 30 minutes, and no 23.1 baseline at 60 minutes. The closest values in the combined 2018/2020 Ohio rows are 10.61 vs 10.70 at 30 min and 20.22 vs 20.49 at 60 min. The abstract also states gains are 'significant', but no statistical tests, confidence intervals, or error bars are reported anywhere in the manuscript. The reader cannot verify or reproduce the headline claims from the tables.
[§IV.E / §III-D.2] The patch configuration is described inconsistently. Section III-D.2 states non-overlapping patches with L_p = 4 and stride L_s = L_p, and adds that L_p = 4 corresponds to 60 minutes at 5-minute sampling (4×5=20, not 60). Section IV.E states 'patch length = 6, stride = 3' for the experiments. These are materially different configurations and would change the architecture and the number of patches. The manuscript must specify exactly which configuration was used and correct the time correspondence, otherwise experiments are not reproducible.
[Table IV ablation] The ablation results do not support the claim that context and retrieval are complementary. In the Ohio 60-minute row, the configuration with RAG but without context (second data row, '✔ ✗ ✔') gives RMSE 20.17, which is lower than the full model's 20.22. This suggests that on this dataset removing context improves RMSE, contradicting the statement that 'the full GlyRAG configuration consistently yields the best or near-best RMSE/MAE.' Additionally, the Table IV BGL-only baseline (21.59) differs from the Table I baseline (20.49), yet both are labeled 'Baseline'/'BGL-only'; the claimed '5% improvement' depends on which baseline is used. The row/column labels (CA, CTL, RAG, Context, BGL) are also not clearly explained, making the ablation hard to interpret.
[§V-A baselines and comparison protocol] The comparison against published results is not controlled. Table I mixes methods that use different input modalities (BGL, I, C, FG, E), different patient splits, and different evaluation protocols, and reports only aggregate RMSE/MAE means without standard deviations. The 'Baseline BGL' row is the authors' own architecture without context/retrieval, not a strong CGM-only baseline such as PatchTST or TimesFM. The abstract's comparison to 'strong CGM-only baselines' and 'PatchTST' is therefore unsupported. At minimum, the authors should provide per-patient error distributions, statistical significance tests, and a direct, controlled comparison against PatchTST on identical splits.

minor comments (5)

[Figure 8 caption] The caption says '1-, 6-, and 12-hour prediction horizons' but the figure and text discuss 5-, 30-, and 60-minute horizons. Please correct the caption or the figure.
[Table III] The values in parentheses in Table III are not labeled; they appear to be the baseline model's values but this is not stated. Add a clear legend and state whether these are from the BGL-only baseline.
[Figure 4] The callout text contains a typo: 'The CMG data' should be 'The CGM data.'
[§IV-C vs §IV-E] The hyperparameter for the translation loss is called λ in Section IV-C but α in Section IV-E and Figure 10. Please use consistent notation.
[Equation (6)] The notation 'arg top-K' is not standard and is not fully precise; consider defining the neighbor set more formally, e.g., the K indices with largest cosine similarity.

Circularity Check

0 steps flagged · score 0.0 of 10

No circular derivation: GlyRAG's forecasting/retrieval pipeline does not reduce to fitted targets or to a self-citation chain; the main problem is a non-circular CGM-only framing issue in the LLM prompt.

full rationale

The paper's forecasting loss (Eq. 1) is a standard Huber regression objective; no parameter is fitted to the evaluation target and then reported as a prediction. The retrieval index D stores (z_j, y_j) pairs, but neighbor selection (Eq. 6) and cross-attention use only the fused embeddings z_jk; the stored target values y_jk are never fed into the attention or MLP, so the RAG module does not leak the future by construction. The cross-translational loss (Eq. 5) is an auxiliary alignment regularizer, not an encoding of the target. The cited prior work by the same group ([11],[15],[25],[34],[39],[40]) is used for datasets, baselines, and related architecture ideas; none is invoked as a uniqueness theorem or as the forcing reason for GlyRAG's design, so the self-citations are not load-bearing. The Limitations section acknowledges noisy context and retrospective evaluation but does not admit a circular step. The genuine concern is non-circular: Section III-C's prompt includes <Carbohydrate Intake>, insulin bolus fields, and asks the LLM for 'Predicted future direction ... next 60 minutes,' which contradicts the abstract's 'CGM-only narrative' claim and means the improvement over the BGL-only baseline could be explained by these extra inputs or by the LLM's own forecast rather than by CGM morphology. That is a validity/leak issue, not a derivation that reduces to its own inputs. Similarly, the abstract's RMSE numbers (13.8->10.6, 23.1->20.2) do not match Table I, but this is a reporting inconsistency, not circularity.

Assumptions & free parameters 5 free parameters · 5 assumptions · 0 invented entities

No new physical/postulated entities; the LLM contextualization agent and RAG module are assembled from existing components. The free parameters are mostly hyperparameters chosen by hand; the key unstated assumption is that the LLM context is CGM-only, which the prompt contradicts.

free parameters (5)

Translation loss weight α = 0.1 (default)
Weight between forecasting and cross-translational loss; authors show sensitivity 0.1–0.5, but the default choice affects ablation conclusions (Table IV, Fig. 10).
Retrieval neighbor count K = 3
Chosen without sensitivity analysis; determines amount of retrieved context in RAG module (Section III-E).
Patch length / stride = L_p=4, L_s=4 (Section III-D.2) vs patch length=6, stride=3 (Section IV-E)
Two inconsistent configurations are given; either choice changes the patch embedding and is part of the model the results depend on.
Huber loss δ = 1.0
Robustness parameter chosen by authors, standard for CGM outliers (Section III-A).
Backbone hyperparameters = d_model=512, n_layers=3, n_heads=4, d_ff=2048, dropout=0.05, LSTM hidden=256
Architecture sizes fixed by authors with no search; they define the baseline and GlyRAG capacity.

assumptions (5)

domain assumption A 3-hour CGM window contains sufficient information to forecast 5/30/60-min horizons
All experiments use L=36 (3h) with no window-length ablation (Section IV-B).
domain assumption LLM-generated qualitative summaries from a frozen GPT-4/LLaMA 3.1 are faithful enough to improve forecasts when embedded by BERT
The LLM is not fine-tuned or validated against clinical expert summaries; authors note this limitation (Section VI).
ad hoc to paper Linear projection between context and glucose embeddings is sufficient to enforce semantic alignment
The cross-translational loss assumes modality A can be linearly translated into modality B (Section III-D.4).
domain assumption Similarity in the learned embedding space corresponds to similar future glucose trajectories
The retrieval module assumes nearest neighbors in fused embedding space are useful case-based analogues (Section III-E).
ad hoc to paper The context is generated from CGM traces alone
The prompt actually includes insulin bolus and carbohydrate intake fields, contradicting the 'CGM-only' framing (Section III-C).

how reviews work

0 comments

Cite this review

Pith. "Pith review of GlyRAG: Context-Aware Retrieval-Augmented Framework for Blood Glucose Forecasting." pith.science (2026). https://pith.science/paper/DJ6G6VQF

@misc{pith2026260105353,
  author       = {Pith},
  title        = {Pith review of: GlyRAG: Context-Aware Retrieval-Augmented Framework for Blood Glucose Forecasting},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/DJ6G6VQF}},
  note         = {Machine review of arXiv:2601.05353}
}

read the original abstract

Accurate blood glucose forecasting using continuous glucose monitoring (CGM) data can support the early prediction of dysglycemic risk. However, current neural-network-based forecasting models treat CGM data as a purely numerical sequence without integrating the contextual information contained in CGM signal morphology. Recently, large language models (LLMs) have shown promise for time-series forecasting, yet their role as agentic context extractors in diabetes care remains largely unexplored. In this study, we bridge glucose forecasting and LLM-based contextualization by developing GlyRAG, a context-aware, retrieval-augmented forecasting framework that uses an LLM as a contextualization agent to summarize glucose morphology directly from a timed CGM window. The generated CGM-only narrative is embedded and fused with patch-based glucose representations, while a retrieval module incorporates similar historical training episodes through cross-attention. We evaluate GlyRAG on the OhioT1DM and AZT1D datasets for 5-, 30-, and 60-minute forecasting horizons. Compared with strong CGM-only baselines, GPT-4 GlyRAG significantly improves long-horizon root mean square error (RMSE) over PatchTST on both datasets. For example, RMSE decreases from 13.8 to 10.6 at 30 minutes and from 23.1 to 20.2 at 60 minutes on OhioT1DM. LLaMA 3.1 produces smaller but significant long-horizon gains, suggesting that the contextualization pipeline is not limited to GPT-4. Clinical error-grid analyses further show that approximately 85 percent of predictions fall within the clinically acceptable Clarke Error Grid Zones A and B. These results suggest that CGM-derived linguistic context and case-based retrieval can improve long-horizon glucose forecasting without requiring additional sensing modalities.

Figures

Figures reproduced from arXiv: 2601.05353 by the authors.

**Figure 2.** Overview of the proposed GlyRAG pipeline: Three-hour CGM windows are extracted and summarized by an LLM agent into short morphology-aware text, which is embedded and fused with glucose patches in a multimodal encoder. A retrieval-augmented module then attends to similar historical patterns to generate multi-horizon forecasts that can be used for behavioral feedback and decision support. user burden, and greater syst… view at source ↗

**Figure 3.** Overall GlyRAG architecture: (a) An LLM agent generates a morphology-aware text summary from the input CGM window, which is encoded by a language model and fused with patch-based glucose embeddings in a multi-head self-attention encoder to produce a joint context–CGM representation (z). (b) The fused query embedding searches a retrieval index for K similar historical episodes; cross-attention branches combine the … view at source ↗

**Figure 4.** Qualitative effect of contextual summaries on glucose forecasting (three examples). Panels (a–c) show a 3-hour CGM window (blue), 12-step/60-min forecasts from GlyRAG (red) and a baseline model (orange), and the ground truth future trajectory (green). Shaded bands indicate clinical ranges (low, target 70–180 mg/dL, high). The callout under each panel is the LLM-generated context summarizing morphology (e.g., postme… view at source ↗

Figures from the paper (6 more)

**Figure 5.** Figure 5: GlyRAG glucose forecasts across prediction horizons (PH = 5, 30, 60 minutes). GlyRAG predictions (red) closely follow actual CGM traces (green) within shaded clinical zones, leveraging contextual morphology to anticipate peaks and nadirs. Dashed oval marks risk markers…

**Figure 6.** Figure 6: Preprocessing and context-extraction workflow: CGM streams are segmented into overlapping 3-h windows with 5/30/60-min targets; an LLM summarizes each window into a context token that is concatenated with the 3-h glucose history to form GlyRAG’s input. B. Preprocessing…

**Figure 7.** Figure 7: Patient-wise RMSE Comparison Across Prediction Horizons for GlyRAG and Baselines. A. Compare with State-of-the-Art (SOTA) Table I summarizes the quantitative comparison in terms of RMSE and MAE at 5-, 30-, and 60-minute prediction horizons. The compared methods employ…

**Figure 8.** Figure 8: Clarke Error Grid analysis for Patients 552 (a) and 570 (b) across 1-, 6-, and 12-hour prediction horizons. Most predictions fall within Zone A, indicating high clinical accuracy, with minor dispersion into Zone B at longer horizons, showing slightly reduced but reliab…

**Figure 9.** Figure 9: Patient-level CG-EGA (point- and rate-error grids) for an Ohio participant (PID 512). datasets. GlyRAG achieves high Accurate Prediction (AP) rates across glucose ranges, particularly in the euglycemic region, while maintaining low Erroneous Prediction (EP) percentage…

**Figure 10.** Figure 10: shows the sensitivity of GlyRAG to the translation-loss weight α, plotting mean RMSE across four patients at 5, 30 and 60-min horizons as α varies from 0.1 0.1 0.2 0.3 0.4 0.5 Alpha ( ) 4 8 12 16 20 RMSE Prediction Horizon 5 min 30 min 60 min [PITH_FULL_IMAGE:figures…

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

42 extracted references · 5 linked inside Pith

[1]

Diabetes,

World Health Organization, “Diabetes,” https://www.who.int/ news-room/fact-sheets/detail/diabetes, 2020, accessed: 2025-01-25

2020
[2]

Intensive diabetes treatment and cardiovascular disease in patients with type 1 diabetes,

D. Control, C. T. of Diabetes Interventions, and C. D. S. R. Group, “Intensive diabetes treatment and cardiovascular disease in patients with type 1 diabetes,”New England Journal of Medicine, vol. 353, no. 25, pp. 2643–2653, 2005

2005
[3]

Incidence of childhood type 1 diabetes worldwide. diabetes mondiale (diamond) project group

M. Karvonen, M. Viik-Kajander, E. Moltchanova, I. Libman, R. LaPorte, and J. Tuomilehto, “Incidence of childhood type 1 diabetes worldwide. diabetes mondiale (diamond) project group.”Diabetes care, vol. 23, no. 10, pp. 1516–1526, 2000

2000
[4]

Idf diabetes atlas,

International Diabetes Federation, “Idf diabetes atlas,” https://www. diabetesatlas.org, 2021, accessed: 2025-01-25

2021
[5]

Intensive glucose control versus conventional glucose control for type 1 diabetes mellitus,

B. Fullerton, K. Jeitler, M. Seitz, K. Horvath, A. Berghold, and A. Siebenhofer, “Intensive glucose control versus conventional glucose control for type 1 diabetes mellitus,”Cochrane Database of Systematic Reviews, no. 2, 2014

2014
[6]

Trajectories of glycemic change in a national cohort of adults with previously controlled type 2 diabetes,

R. G. McCoy, C. Ngufor, H. K. Van Houten, B. Caffo, and N. D. Shah, “Trajectories of glycemic change in a national cohort of adults with previously controlled type 2 diabetes,”Medical care, vol. 55, no. 11, pp. 956–964, 2017

2017
[7]

Glucose sensor with predictive alarm for hypoglycaemia: Improved glycaemic control in adolescents with type 1 diabetes,

M. Marigliano, C. Piona, V . Mancioppi, E. Morotti, A. Morandi, and C. Maffeis, “Glucose sensor with predictive alarm for hypoglycaemia: Improved glycaemic control in adolescents with type 1 diabetes,” Diabetes, Obesity and Metabolism, vol. 26, no. 4, pp. 1314–1320, 2024

2024
[8]

Data-driven modeling and prediction of blood glucose dynamics: Machine learning applications in type 1 diabetes,

A. Z. Woldaregay, E. ˚Arsand, S. Walderhaug, D. Albers, L. Mamykina, T. Botsis, and G. Hartvigsen, “Data-driven modeling and prediction of blood glucose dynamics: Machine learning applications in type 1 diabetes,”Artificial intelligence in medicine, vol. 98, pp. 109–134, 2019

2019

Show all 42 references

[9]

Advances in continuous glucose moni- toring: clinical applications,

S. Y . Kwon and J. S. Moon, “Advances in continuous glucose moni- toring: clinical applications,”Endocrinology and Metabolism, vol. 40, no. 2, pp. 161–173, 2025

2025
[10]

Continuous glucose monitoring sensors for diabetes management: a review of technologies and applications,

G. Cappon, M. Vettoretti, G. Sparacino, and A. Facchinetti, “Continuous glucose monitoring sensors for diabetes management: a review of technologies and applications,”Diabetes & metabolism journal, vol. 43, no. 4, p. 383, 2019

2019
[11]

Glytwin: Digital twin for glucose control in type 1 diabetes through optimal behavioral modifications using patient-centric counterfactuals,

A. Arefeen, S. Khamesian, M. A. Grando, B. Thompson, and H. Ghasemzadeh, “Glytwin: Digital twin for glucose control in type 1 diabetes through optimal behavioral modifications using patient-centric counterfactuals,” 2025. [Online]. Available: https: //arxiv.org/abs/2504.09846

2025 arXiv
[12]

Benchmarking machine learning algorithms on blood glucose prediction for type i diabetes in comparison with classical time-series models,

J. Xie and Q. Wang, “Benchmarking machine learning algorithms on blood glucose prediction for type i diabetes in comparison with classical time-series models,”IEEE Transactions on Biomedical Engineering, vol. 67, no. 11, pp. 3101–3124, 2020

2020
[13]

Machine learning techniques for hypoglycemia prediction: Trends and challenges,

O. Mujahid, I. Contreras, and J. Vehi, “Machine learning techniques for hypoglycemia prediction: Trends and challenges,”Sensors, vol. 21, no. 2, p. 546, 2021

2021
[14]

Generalized multi task learning framework for glucose forecasting and hypoglycemia detection using simulation to reality,

M. Hwang, V . P. Rachim, J. Yoo, Y . Lee, and S.-M. Park, “Generalized multi task learning framework for glucose forecasting and hypoglycemia detection using simulation to reality,”npj Digital Medicine, vol. 8, no. 1, p. 612, 2025

2025
[15]

Time-aware cross-attention for multi-modal sensor-based blood glucose forecasting,

A. Machiraju, E. Farahmand, S. B. Soumma, A. Arefeen, C. Johnston, and H. Ghasemzadeh, “Time-aware cross-attention for multi-modal sensor-based blood glucose forecasting,” inIEEE-EMBS International Conference on Body Sensor Networks 2025, 2025. [Online]. Available: https://ope...

2025
[16]

Glunet: A deep learning framework for accurate glucose forecasting,

K. Li, C. Liu, T. Zhu, P. Herrero, and P. Georgiou, “Glunet: A deep learning framework for accurate glucose forecasting,”IEEE journal of biomedical and health informatics, vol. 24, no. 2, pp. 414–423, 2019

2019
[17]

Deep multitask learning by stacked long short-term memory for predicting personalized blood glucose concentration,

M. M. H. Shuvo and S. K. Islam, “Deep multitask learning by stacked long short-term memory for predicting personalized blood glucose concentration,”IEEE Journal of Biomedical and Health Informatics, vol. 27, no. 3, pp. 1612–1623, 2023

2023
[18]

A decoder-only foundation model for time-series forecasting,

A. Das, W. Kong, R. Sen, and Y . Zhou, “A decoder-only foundation model for time-series forecasting,” inForty-first International Confer- ence on Machine Learning, 2024

2024
[19]

Time-LLM: Time series forecasting by reprogramming large language models,

M. Jin, S. Wang, L. Ma, Z. Chu, J. Y . Zhang, X. Shi, P.-Y . Chen, Y . Liang, Y .-F. Li, S. Pan, and Q. Wen, “Time-LLM: Time series forecasting by reprogramming large language models,” inThe Twelfth International Conference on Learning Representations, 2024. [Online]. Availabl...

2024
[20]

A deep learning approach for blood glucose prediction of type 1 diabetes,

J. Freiburghaus, A. Rizzotti, and F. Albertetti, “A deep learning approach for blood glucose prediction of type 1 diabetes,” inProceedings of the Proceedings of the 5th International Workshop on Knowledge Discovery in Healthcare Data co-located with 24th European Conference on...

2020
[21]

Long-term prediction of blood glucose levels in type 1 diabetes using a cnn-lstm-based deep neural network,

M. Jaloli and M. Cescon, “Long-term prediction of blood glucose levels in type 1 diabetes using a cnn-lstm-based deep neural network,”Journal of diabetes science and technology, vol. 17, no. 6, pp. 1590–1601, 2023

2023
[22]

Integration of clin- ical criteria into the training of deep models: Application to glucose prediction for diabetic people,

M. De Bois, M. A. El-Yacoubi, and M. Ammi, “Integration of clin- ical criteria into the training of deep models: Application to glucose prediction for diabetic people,”Smart Health, vol. 21, p. 100193, 2021

2021
[23]

The importance of interpreting machine learning models for blood glucose prediction in diabetes: an analysis using shap,

F. Prendin, J. Pavan, G. Cappon, S. Del Favero, G. Sparacino, and A. Facchinetti, “The importance of interpreting machine learning models for blood glucose prediction in diabetes: an analysis using shap,” Scientific reports, vol. 13, no. 1, p. 16865, 2023

2023
[24]

A multitask learning approach to personalized blood glucose prediction,

J. Daniels, P. Herrero, and P. Georgiou, “A multitask learning approach to personalized blood glucose prediction,”IEEE Journal of Biomedical and Health Informatics, vol. 26, no. 1, pp. 436–445, 2021

2021
[25]

Hybrid attention model using feature decomposition and knowledge distillation for glucose forecasting,

E. Farahmand, S. B. Soumma, N. T. Chatrudi, and H. Ghasemzadeh, “Hybrid attention model using feature decomposition and knowledge distillation for glucose forecasting,” 2025. [Online]. Available: https://arxiv.org/abs/2411.10703

2025 arXiv
[26]

Timegpt-1,

A. Garza and M. Mergenthaler-Canseco, “Timegpt-1,” 2023

2023
[27]

Promptcast: A new prompt-based learning paradigm for time series forecasting,

H. Xue and F. D. Salim, “Promptcast: A new prompt-based learning paradigm for time series forecasting,”IEEE Transactions on Knowledge and Data Engineering, vol. 36, no. 11, pp. 6851–6864, 2023

2023
[28]

Large language models are few-shot health learners,

X. Liu, D. McDuff, G. Kovacs, I. Galatzer-Levy, J. Sunshine, J. Zhan, M.-Z. Poh, S. Liao, P. Di Achille, and S. Patel, “Large language models are few-shot health learners,”arXiv preprint arXiv:2305.15525, 2023

2023 arXiv
[29]

Timecap: Learning to contextualize, augment, and predict time series events with large language model agents,

G. Lee, W. Yu, K. Shin, W. Cheng, and H. Chen, “Timecap: Learning to contextualize, augment, and predict time series events with large language model agents,” inProceedings of the AAAI Conference on Artificial Intelligence, vol. 39, no. 17, 2025, pp. 18 082–18 090

2025
[30]

Generalized huber loss for robust learning and its efficient minimization for a robust statistics,

K. Gokcesu and H. Gokcesu, “Generalized huber loss for robust learning and its efficient minimization for a robust statistics,” 2021. [Online]. Available: https://arxiv.org/abs/2108.12627

2021 arXiv
[31]

Bert: Pre-training of deep bidirectional transformers for language understanding,

J. Devlin, M.-W. Chang, K. Lee, and K. Toutanova, “Bert: Pre-training of deep bidirectional transformers for language understanding,” inNorth American Chapter of the Association for Computational Linguistics,
[32]

A time series is worth 64 words: Long-term forecasting with transformers,

Y . Nie, N. H. Nguyen, P. Sinthong, and J. Kalagnanam, “A time series is worth 64 words: Long-term forecasting with transformers,” in International Conference on Learning Representations, 2023

2023
[33]

The ohiot1dm dataset for blood glucose level prediction: Update 2020,

C. Marling and R. Bunescu, “The ohiot1dm dataset for blood glucose level prediction: Update 2020,” inCEUR workshop proceedings, vol. 2675, 2020, p. 71

2020
[34]

Azt1d: A real-world dataset for type 1 diabetes,

S. Khamesian, A. Arefeen, B. M. Thompson, A. Grando, and H. Ghasemzadeh, “Azt1d: A real-world dataset for type 1 diabetes,” 2025. [Online]. Available: https://data.mendeley.com/datasets/ gk9m674wcx/1

2025
[35]

Blood glucose prediction with variance estimation using recurrent neural networks,

J. Martinsson, A. Schliep, B. Eliasson, and O. Mogren, “Blood glucose prediction with variance estimation using recurrent neural networks,” Journal of Healthcare Informatics Research, vol. 4, pp. 1–18, 2020

2020
[36]

Deep residual time-series forecasting: Application to blood glucose prediction

H. Rubin-Falcone, I. Fox, and J. Wiens, “Deep residual time-series forecasting: Application to blood glucose prediction.”KDH@ ECAI, vol. 20, pp. 105–109, 2020

2020
[37]

Investigating potentials and pitfalls of knowledge distillation across datasets for blood glucose forecasting,

H. Hameed and S. Kleinberg, “Investigating potentials and pitfalls of knowledge distillation across datasets for blood glucose forecasting,” in Proceedings of the 5th Annual Workshop on Knowledge Discovery in Healthcare Data, 2020

2020
[38]

Personalised short-term glucose prediction via recurrent self-attention network,

R. Cui, C. Hettiarachchi, C. J. Nolan, E. Daskalaki, and H. Suominen, “Personalised short-term glucose prediction via recurrent self-attention network,” in2021 IEEE 34th International Symposium on Computer- Based Medical Systems (CBMS). IEEE, 2021, pp. 154–159

2021
[39]

Glysim: Modeling and simulating glycemic response for behavioral lifestyle interventions,

A. Arefeen and H. Ghasemzadeh, “Glysim: Modeling and simulating glycemic response for behavioral lifestyle interventions,” in2023 IEEE EMBS International Conference on Biomedical and Health Informatics (BHI). IEEE, 2023, pp. 1–5

2023
[40]

Type 1 diabetes management using glimmer: Glucose level indicator model with modified error rate,

S. Khamesian, A. Arefeen, M. A. Grando, B. M. Thompson, and H. Ghasemzadeh, “Type 1 diabetes management using glimmer: Glucose level indicator model with modified error rate,” 2025. [Online]. Available: https://arxiv.org/abs/2502.14183

2025 arXiv
[41]

Evaluating clinical accuracy of systems for self-monitoring of blood glucose,

W. L. Clarke, D. Cox, L. A. Gonder-Frederick, W. Carter, and S. L. Pohl, “Evaluating clinical accuracy of systems for self-monitoring of blood glucose,”Diabetes Care, vol. 10, no. 5, pp. 622–628, 09 1987. [Online]. Available: https://doi.org/10.2337/diacare.10.5.622 Shovito Ba...

1987 doi
[2019]

Available: https://api.semanticscholar.org/CorpusID: 52967399

[Online]. Available: https://api.semanticscholar.org/CorpusID: 52967399

Pith tools

Reviewed August 3, 2026 · model on record in the stance chip above.

[1] [1]

Diabetes,

World Health Organization, “Diabetes,” https://www.who.int/ news-room/fact-sheets/detail/diabetes, 2020, accessed: 2025-01-25

2020

[2] [2]

Intensive diabetes treatment and cardiovascular disease in patients with type 1 diabetes,

D. Control, C. T. of Diabetes Interventions, and C. D. S. R. Group, “Intensive diabetes treatment and cardiovascular disease in patients with type 1 diabetes,”New England Journal of Medicine, vol. 353, no. 25, pp. 2643–2653, 2005

2005

[3] [3]

Incidence of childhood type 1 diabetes worldwide. diabetes mondiale (diamond) project group

M. Karvonen, M. Viik-Kajander, E. Moltchanova, I. Libman, R. LaPorte, and J. Tuomilehto, “Incidence of childhood type 1 diabetes worldwide. diabetes mondiale (diamond) project group.”Diabetes care, vol. 23, no. 10, pp. 1516–1526, 2000

2000

[4] [4]

Idf diabetes atlas,

International Diabetes Federation, “Idf diabetes atlas,” https://www. diabetesatlas.org, 2021, accessed: 2025-01-25

2021

[5] [5]

Intensive glucose control versus conventional glucose control for type 1 diabetes mellitus,

B. Fullerton, K. Jeitler, M. Seitz, K. Horvath, A. Berghold, and A. Siebenhofer, “Intensive glucose control versus conventional glucose control for type 1 diabetes mellitus,”Cochrane Database of Systematic Reviews, no. 2, 2014

2014

[6] [6]

Trajectories of glycemic change in a national cohort of adults with previously controlled type 2 diabetes,

R. G. McCoy, C. Ngufor, H. K. Van Houten, B. Caffo, and N. D. Shah, “Trajectories of glycemic change in a national cohort of adults with previously controlled type 2 diabetes,”Medical care, vol. 55, no. 11, pp. 956–964, 2017

2017

[7] [7]

Glucose sensor with predictive alarm for hypoglycaemia: Improved glycaemic control in adolescents with type 1 diabetes,

M. Marigliano, C. Piona, V . Mancioppi, E. Morotti, A. Morandi, and C. Maffeis, “Glucose sensor with predictive alarm for hypoglycaemia: Improved glycaemic control in adolescents with type 1 diabetes,” Diabetes, Obesity and Metabolism, vol. 26, no. 4, pp. 1314–1320, 2024

2024

[8] [8]

Data-driven modeling and prediction of blood glucose dynamics: Machine learning applications in type 1 diabetes,

A. Z. Woldaregay, E. ˚Arsand, S. Walderhaug, D. Albers, L. Mamykina, T. Botsis, and G. Hartvigsen, “Data-driven modeling and prediction of blood glucose dynamics: Machine learning applications in type 1 diabetes,”Artificial intelligence in medicine, vol. 98, pp. 109–134, 2019

2019