Assessing Foundation Models' Transferability to Physiological Signals in Precision Medicine

Brian Locke; Cove Geary; Matthias Christenson; Pranav Koirala; Warren Woodrich Pettine

REVIEW 4 major objections 5 minor 16 references

Assessing Foundation Models' Transferability to Physiological Signals in Precision Medicine

T0 review · 4 major / 5 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read A general-purpose time series foundation model, Moirai, produces embeddings that systematically distort physiological signals, adding spurious cross-feature correlations, destroying temporal smoothness, and reducing the ability to decode…

desk verdict Useful evaluation pipeline, but the per-feature projection confounds the negative result; worth a serious round of revisions. read the letter →

arxiv 2412.03427 v1 pith:VIANZJQR submitted 2024-12-04 cs.LG

Matthias Christenson , Cove Geary , Brian Locke , Pranav Koirala , Warren Woodrich Pettine This is my paper · ORCID

classification cs.LG

keywords foundationmodelstimeseriesphysiologicalsignalsprecisionmedicinetransferlearningembeddingssyntheticphysiologysimulationfeatureentanglement

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

This paper builds a three-stage pipeline to test whether general-purpose time series foundation models can be used for precision medicine: generate clinically relevant synthetic physiological signals, push them through the model's transformer to get embeddings, and probe the embeddings with linear methods. Applying the pipeline to the Moirai time series foundation model, the paper finds that the embeddings are not faithful representations of physiological data. Features that are nearly independent in the raw signals become correlated in embedding space, raw signals cannot be reconstructed from embeddings, temporal smoothness and dimensionality drop, and medical scenarios become harder to tell apart. A logistic-regression probe decodes feature identity from raw signals with mean AUC 0.96 but only 0.78 from embeddings. The paper concludes that general-purpose time series foundation models will need architectural changes or targeted fine-tuning before clinical deployment.

What carries the argument

The machinery is a three-stage evaluation pipeline. Stage one uses whole-body physiology simulations to generate diverse, clinically relevant scenarios such as hemorrhage, sepsis, multi-organ failure, burn, and radiation exposure, with signals normalized to zero mean and unit variance. Stage two linearly interpolates each feature to 1000 timesteps and passes each feature independently through the foundation model's transformer module, taking the final-layer embedding. Stage three applies linear probes: Pearson correlations between feature pairs, linear regression with 5-fold cross-validation for reconstruction R2, PCA for dimensionality and a smoothness metric based on velocity ratios relative to randomly permuted trajectories, cosine similarity between scenario embeddings, and logistic regression AUC for pairwise feature identity decoding. These probes are designed to quantify feature entanglement, temporal dynamics preservation, and medical scenario differentiation.

What would settle it

An experiment that feeds the same physiological scenarios as joint multivariate tensors with native patch sizes and then measures the same metrics would settle the question: if the distortions vanish, the paper's conclusion fails; if they persist, the conclusion is robust.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central discovery is that feeding each physiological feature independently through Moirai's transformer module produces final-layer embeddings that systematically corrupt the signal structure: spurious cross-feature correlations appear, linear reconstruction of the raw time series from embeddings is poor, the low-dimensional PCA trajectories become erratic and lower-dimensional, inter-scenario correlation rises, and pairwise feature decoding drops from near-perfect on raw signals to 0.78 AUC on embeddings. These are measured on synthetic physiological signals from BioGears covering hemorrhage, sepsis, multi-organ failure, and related scenarios. The paper positions this as evidence that zero-shot transfer of such foundation models to precision medicine is not ready, and that the evaluation pipeline itself is the contribution that makes such assessments fast and inexpensive.

Load-bearing premise

The pipeline's verdict depends on the assumption that the way signals were fed to the model—one feature at a time, stretched to a fixed length of 1000 time steps—is a fair test rather than an awkward input format the model was not built for.

Editorial extensions

If this is right

If Moirai-level distortions generalize to other time series foundation models, zero-shot use of such embeddings in clinical pipelines will misrepresent physiological state and can mislead downstream models.
The pipeline's simulation-to-linear-probe design makes it possible to benchmark many foundation models quickly and cheaply before any clinical data collection.
The identified failure modes each suggest a fine-tuning objective: controlled-correlation scenarios to combat feature entanglement, scenarios with clinically relevant timescales to restore temporal dynamics, and edge-case scenario design to improve discrimination.
Feature decoding AUC dropping from 0.96 to 0.78 implies that subtle but clinically important features such as central venous pressure and renal blood flow will be the first to be confounded.
Since no fine-tuning was performed, the results bound zero-shot performance, not the ceiling after adaptation.

Reading between the lines

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

A testable extension the paper does not run: feed the same physiological scenarios as joint multivariate channels or vary patch size and interpolation length to see how much of the distortion is due to the independent-feature projection rather than the model itself.
If feature entanglement in embedding space is as strong as reported, a simple linear whitening or decorrelation post-processing step might recover some feature independence, but the reconstruction and temporal-smoothness losses imply information is genuinely discarded, so post-hoc correction alone is unlikely to suffice.
The same evaluation could be applied to wearable photoplethysmography or ECG streams, where temporal smoothness matters for beat-by-beat analysis, to see whether the distortion pattern persists across physiological signal types.

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

Colleague,

The paper proposes a simulation-to-linear-probe pipeline for checking whether a time series foundation model preserves physiological signal structure, and demonstrates it on Moirai. The pipeline idea is genuinely useful: BioGears gives you controllable, clinically themed scenarios, and the linear probing is cheap. If you're a group deciding whether to fine-tune a foundation model for monitoring data, this kind of capability audit is worth having. The paper's negative result is also not trivial: Moirai's embeddings do show feature entanglement, poor reconstruction, distorted temporal dynamics, and worse scenario discrimination in their protocol. That's a plausible red flag.

The soft spots are real, though. The biggest one is the projection step in Section 2.2: each physiological feature is interpolated and passed through Moirai independently. So the model never sees a multivariate physiological state; it sees seven unrelated univariate series. Under that protocol, the spurious cross-feature correlations could be an artifact of how the model handles input formatting and patching, not an intrinsic property of the representations. The authors never show a control where the features are fed jointly through the model's native multivariate interface, so the central conclusion — that 'current foundation models may require substantial architectural modifications' — overreaches. Also missing: code/data, simulation counts, error bars on most metrics, any real-data validation. The AUC threshold of 0.9 for 'effective disentanglement' is presented without justification. And there's a concrete inconsistency: the text lists acute hemorrhage, sepsis, and multi-organ failure, but Figure 1 shows a burn scenario and a radiation scenario. That needs fixing.

I'm not saying the result is fake; the degradation is consistently reported and the paper is honest about being work-in-progress. But the load-bearing claim about what Moirai does to physiological signals is confounded by the evaluation protocol. The good news is that the confound is addressable with a small number of additional experiments. I'd send this to a serious referee — the pipeline is useful enough and the negative result is relevant to the time series foundation model community — but I'd flag the missing multivariate control as a major revision. If the authors add that control, tighten the conclusions, and release the code, the paper becomes a solid contribution.

Referee Report

4 major / 5 minor

Summary. The paper presents a work-in-progress pipeline for assessing how well time-series foundation models transfer to physiological signals in precision medicine. The pipeline generates synthetic physiological data with the BioGears simulator across several clinical scenarios (e.g., hemorrhage, sepsis, multi-organ failure), linearly interpolates each of seven physiological features to 1000 timesteps, independently projects each feature through the Moirai time-series foundation model to obtain final-layer embeddings, and then evaluates those embeddings with linear probes: Pearson correlations among features, linear reconstruction R², PCA-based dimensionality and trajectory smoothness, cosine similarity between scenarios, and logistic-regression decoding of feature identity. The authors report that Moirai's embeddings introduce spurious cross-feature correlations, poorly reconstruct the raw signals, destroy temporal smoothness and dimensionality, reduce scenario discrimination, and lower feature-decoding AUC from 0.96 to 0.78. From these results they conclude that current foundation models may require substantial architectural modification or targeted fine-tuning before clinical deployment.

Significance. If the reported effects were shown to be intrinsic to the model class, the paper would provide a useful, low-cost benchmarking methodology for an important application area. The pipeline is clearly described and the use of linear probes is appropriate for isolating representation quality from classifier power. The authors also deserve credit for explicitly framing the work as a work-in-progress and for naming ongoing extensions. However, the central claim as stated is not yet supported: the evaluation protocol projects each feature independently, so the reported distortions may be artifacts of input formatting rather than intrinsic limitations of the foundation model. The paper also provides no code, no data, no real-physiological-signal validation, and reports almost no error bars or experimental scale, which limits the reproducibility and robustness of the quantitative findings. The contribution is therefore a promising evaluation framework with initial, but not yet conclusive, results.

major comments (4)

[§2.2] The evaluation rests on a per-feature projection: each of the seven features is linearly interpolated to 1000 timesteps and passed independently through Moirai's transformer module, so the model never sees the multivariate physiological state. Under this protocol, the reported spurious cross-feature correlations (§3.2), reconstruction loss (§3.3), and temporal dynamics distortions (§3.4) could be artifacts of formatting and patch construction for univariate series rather than intrinsic limitations of the model's representation of physiological signals. A mandatory control is to feed the same signals jointly through Moirai's native multivariate interface, or at least to vary patch size and interpolation length and to compare against a baseline encoder; without such a control, the central conclusion in §4.2 is not supported.
[§3.6 and §2.3] The feature-decoding experiment is interpreted with an arbitrary threshold: AUC > 0.9 is called 'effective disentanglement,' and the drop from raw AUC 0.96 to embedded AUC 0.78 is presented as compromised clinical utility. But pairwise linear separability of feature identity is a measure of feature distinctiveness, not of disentanglement, and a 0.78 AUC is not by itself evidence of clinical non-usability. The paper does not tie decoding performance to any downstream clinical task, so the claim that 'reduced capacity for feature decoding impairs clinical utility' is not established.
[§3] The quantitative results are reported without the experimental scale needed to assess them: the number of BioGears virtual patients, scenarios, sessions, and simulation runs is never stated, and only the decoding AUC has error bars. The smoothness metric depends on 1000 random permutations, but no sampling distribution or sensitivity analysis is given; the PCA variance threshold (90%) and the decoding threshold (0.9) are fixed without justification. These omissions make it impossible to know whether the reported distortions are robust or within noise.
[§4.1 and §4.2] The scope of the empirical evidence is one model (Moirai), one simulation package's default scenarios, and no real physiological recordings. The abstract and §4.2 nevertheless conclude that 'current foundation models may require substantial architectural modifications or targeted fine-tuning before deployment in clinical settings.' This overgeneralizes from a single univariate projection protocol; at minimum, the claims should be restricted to Moirai's zero-shot univariate embeddings and to the BioGears synthetic distribution.

minor comments (5)

[§3.2–§3.5] The text repeatedly refers to 'Fig. 2.3' where the only pipeline figure is labeled Fig. 1; all cross-references should be corrected.
[Throughout] The model name is spelled inconsistently as both 'Moirai' and 'Moiria' (for example in the abstract, §4.1, §4.2, and the Fig. 1 caption); the latter is a typo.
[§2.2] The linear interpolation to 1000 timesteps is described but its effect is never analyzed; a sensitivity analysis over interpolation lengths would strengthen the claim that the results are not formatting artifacts.
[§2.3 and §3.6] The logistic-regression decoding setup is underspecified: which feature pairs, how many samples, and which regularization were used are not reported, and the pairwise AUC values for particular features mentioned in §3.6 are not shown.
[Fig. 1] The figure contains many panels with dense small text; the caption does not identify the simulated scenario shown in panels D(left) and E(top-left), and the color and type legends are hard to distinguish in print.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the evaluation empirically compares external model embeddings against raw simulated signals with independent metrics.

full rationale

The paper presents an empirical evaluation pipeline rather than a derivation. Raw physiological signals are generated by BioGears, an external simulation package, and embeddings are produced by Moirai, an external foundation model; no fitted parameter in the paper is later reported as a prediction. The reported metrics compare raw and embedded representations using distinct operationalizations: feature correlation uses Pearson correlation, reconstruction uses 5-fold cross-validated linear regression R2, temporal dynamics uses PCA dimensionality and a permutation-normalized smoothness measure, scenario separation uses cosine similarity and PCA, and feature decoding uses logistic-regression AUC on held-out data. None of these quantities is defined in terms of another result of the paper, and none is forced by construction to produce the reported direction of the effects. The paper contains no load-bearing self-citations: references [2] and [14] are independent external works for the simulator and the model. The concern that feeding each feature independently through the transformer may make the distortions artifacts of input formatting is a legitimate question about experimental design and external validity, but it is not circularity, because the conclusion is not true by definition of the metrics or by the paper's own equations. The empirical comparison is self-contained against external benchmarks, so the appropriate circularity score is 0.

Assumptions & free parameters 4 free parameters · 5 assumptions · 0 invented entities

The evaluation depends on synthetic data fidelity, input formatting, and chosen thresholds. None of these is externally validated, so the central claim rests on several unverified domain assumptions rather than on fitted model parameters.

free parameters (4)

Interpolation sequence length = 1000 timesteps
Hand-chosen in Section 2.2; no sensitivity analysis is provided.
PCA variance threshold = 90%
Hand-chosen in Section 2.3 for dimensionality and smoothness metrics.
AUC threshold for effective disentanglement = 0.9
Hand-chosen criterion in Section 2.3; values below are labeled entangled.
Number of random permutations for smoothness = 1000
Hand-chosen in Section 2.3 for the null smoothness baseline.

assumptions (5)

domain assumption BioGears default scenarios produce physiological signals whose statistical structure is representative of real clinical precision medicine data.
The entire evaluation and the conclusion about clinical deployment rest on synthetic data standing in for real patient physiology; no real-data validation is provided (Section 2.1, Discussion Section 4.1).
domain assumption Linear interpolation to a common 1000-timestep grid preserves the temporal dynamics and feature relationships of the original signals.
All embeddings and raw comparisons use the interpolated version; no sensitivity analysis to interpolation length is reported (Section 2.2).
domain assumption Passing each physiological feature independently through the transformer module yields a valid embedding of the multivariate physiological state.
Crucial modeling choice in Section 2.2; the paper does not compare with multivariate or joint-input embeddings, so observed distortions could be artifacts of the input format.
domain assumption PCA dimensionality and trajectory smoothness of the raw signals are appropriate ground-truth descriptors of physiological dynamics.
The metrics in Section 2.3 treat raw signal smoothness and high dimensionality as desirable without establishing these properties as clinically necessary.
ad hoc to paper Moirai is a representative example of current time series foundation models.
Only one model is tested; the paper itself states extension to multiple architectures is ongoing (Section 4.1).

how reviews work

0 comments

Cite this review

Pith. "Pith review of Assessing Foundation Models' Transferability to Physiological Signals in Precision Medicine." pith.science (2026). https://pith.science/paper/VIANZJQR

@misc{pith2026241203427,
  author       = {Pith},
  title        = {Pith review of: Assessing Foundation Models' Transferability to Physiological Signals in Precision Medicine},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/VIANZJQR}},
  note         = {Machine review of arXiv:2412.03427}
}

read the original abstract

The success of precision medicine requires computational models that can effectively process and interpret diverse physiological signals across heterogeneous patient populations. While foundation models have demonstrated remarkable transfer capabilities across various domains, their effectiveness in handling individual-specific physiological signals - crucial for precision medicine - remains largely unexplored. This work introduces a systematic pipeline for rapidly and efficiently evaluating foundation models' transfer capabilities in medical contexts. Our pipeline employs a three-stage approach. First, it leverages physiological simulation software to generate diverse, clinically relevant scenarios, particularly focusing on data-scarce medical conditions. This simulation-based approach enables both targeted capability assessment and subsequent model fine-tuning. Second, the pipeline projects these simulated signals through the foundation model to obtain embeddings, which are then evaluated using linear methods. This evaluation quantifies the model's ability to capture three critical aspects: physiological feature independence, temporal dynamics preservation, and medical scenario differentiation. Finally, the pipeline validates these representations through specific downstream medical tasks. Initial testing of our pipeline on the Moirai time series foundation model revealed significant limitations in physiological signal processing, including feature entanglement, temporal dynamics distortion, and reduced scenario discrimination. These findings suggest that current foundation models may require substantial architectural modifications or targeted fine-tuning before deployment in clinical settings.

Figures

Figures reproduced from arXiv: 2412.03427 by the authors.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

16 extracted references · 9 canonical work pages

[1]

A survey on data-efficient algorithms in big data era.Journal of Big Data, 8(1):24, January 2021

Amina Adadi. A survey on data-efficient algorithms in big data era.Journal of Big Data, 8(1):24, January 2021

work page 2021
[2]

BioGears: A C++ library for whole body physiology simulations.Journal of Open Source Software, 5(56):2645, December 2020

AustinBaird,MatthewMcDaniel,StevenWhite,NathanTatum,andLucasMarin. BioGears: A C++ library for whole body physiology simulations.Journal of Open Source Software, 5(56):2645, December 2020

work page 2020
[3]

Ana Barragán-Montero, Umair Javaid, Gilmer Valdés, Dan Nguyen, Paul Desbor- des, Benoit Macq, Siri Willems, Liesbeth Vandewinckele, Mats Holmström, Fredrik Löfman, Steven Michiels, Kevin Souris, Edmond Sterpin, and John A. Lee. Arti- ficial intelligence and machine learning for medical imaging: A technology review. Physica Medica, 83:242–256, March 2021

work page 2021
[4]

Hudson, Ehsan Adeli, Russ Altman, Simran Arora, Sydney von Arx, Michael S

Rishi Bommasani, Drew A. Hudson, Ehsan Adeli, Russ Altman, Simran Arora, Sydney von Arx, Michael S. Bernstein, Jeannette Bohg, Antoine Bosselut, Emma Brunskill, Erik Brynjolfsson, Shyamal Buch, Dallas Card, Rodrigo Castellon, Ni- ladri Chatterji, Annie Chen, Kathleen Creel, Jared Quincy Davis, Dora Dem- szky, Chris Donahue, Moussa Doumbouya, Esin Durmus, ...

arXiv 2022
[5]

Sparks of Artificial General Intelligence: Early experiments with GPT-4, April 2023

Sébastien Bubeck, Varun Chandrasekaran, Ronen Eldan, Johannes Gehrke, Eric Horvitz, Ece Kamar, Peter Lee, Yin Tat Lee, Yuanzhi Li, Scott Lundberg, Harsha Nori, Hamid Palangi, Marco Tulio Ribeiro, and Yi Zhang. Sparks of Artificial General Intelligence: Early experiments with GPT-4, April 2023. arXiv:2303.12712 [cs]

arXiv 2023
[6]

Theis, Alan Moses, and Bo Wang

MicaelaE.Consens,CameronDufault,MichaelWainberg,DuncanForster,Mehran Karimzadeh, Hani Goodarzi, Fabian J. Theis, Alan Moses, and Bo Wang. To Transformers and Beyond: Large Language Models for the Genome, November

work page
[7]

The MIT Press, Cambridge, Mass, November 2016

Ian Goodfellow, Yoshua Bengio, and Aaron Courville.Deep Learning. The MIT Press, Cambridge, Mass, November 2016

work page 2016
[8]

MOMENT: A Family of Open Time-series Foundation Models, February 2024

Mononito Goswami, Konrad Szafer, Arjun Choudhry, Yifu Cai, Shuo Li, and Ar- tur Dubrawski. MOMENT: A Family of Open Time-series Foundation Models, February 2024. arXiv:2402.03885 [cs]

arXiv 2024

Show all 16 references

[9]

Andrew Taylor, Harlan M

Irene Li, Jessica Pan, Jeremy Goldwasser, Neha Verma, Wai Pan Wong, Muhammed Yavuz Nuzumlalı, Benjamin Rosand, Yixin Li, Matthew Zhang, David Chang, R. Andrew Taylor, Harlan M. Krumholz, and Dragomir Radev. Neural Natural Language Processing for unstructured data in electronic...

2022
[10]

Foundation Models for Time Series Analysis: A Tutorial and Survey, April 2024

Yuxuan Liang, Haomin Wen, Yuqi Nie, Yushan Jiang, Ming Jin, Dongjin Song, Shirui Pan, and Qingsong Wen. Foundation Models for Time Series Analysis: A Tutorial and Survey, April 2024. arXiv:2403.14735 [cs]

2024 arXiv
[11]

Henry T. Peng, M. Musaab Siddiqui, Shawn G. Rhind, Jing Zhang, Luis Teodoro da Luz, and Andrew Beckett. Artificial intelligence and machine learning for hem- orrhagic trauma care. Military Medical Research, 10(1):6, February 2023

2023
[12]

Khaled Saab, Tao Tu, Wei-Hung Weng, Ryutaro Tanno, David Stutz, Ellery Wul- czyn, Fan Zhang, Tim Strother, Chunjong Park, Elahe Vedadi, Juanma Zambrano Chaves, Szu-Yeu Hu, Mike Schaekermann, Aishwarya Kamath, Yong Cheng, David G. T. Barrett, Cathy Cheung, Basil Mustafa, Anil P...

2024 arXiv
[13]

Teng and Adam B

Andrew K. Teng and Adam B. Wilcox. A Review of Predictive Analytics Solu- tions for Sepsis Patients. Applied Clinical Informatics , 11(3):387–398, May 2020. Publisher: Georg Thieme Verlag KG

2020
[14]

Unified Training of Universal Time Series Forecasting Transformers, February 2024

Gerald Woo, Chenghao Liu, Akshat Kumar, Caiming Xiong, Silvio Savarese, and Doyen Sahoo. Unified Training of Universal Time Series Forecasting Transformers, February 2024. arXiv:2402.02592 [cs]

2024 arXiv
[15]

L3MVN: Leveraging Large Lan- guage Models for Visual Target Navigation

Bangguo Yu, Hamidreza Kasaei, and Ming Cao. L3MVN: Leveraging Large Lan- guage Models for Visual Target Navigation. In 2023 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) , pages 3554–3560, October

2023
[2023]

arXiv:2311.07621 [cs, q-bio]

arXiv

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.

[1] [1]

A survey on data-efficient algorithms in big data era.Journal of Big Data, 8(1):24, January 2021

Amina Adadi. A survey on data-efficient algorithms in big data era.Journal of Big Data, 8(1):24, January 2021

work page 2021

[2] [2]

BioGears: A C++ library for whole body physiology simulations.Journal of Open Source Software, 5(56):2645, December 2020

AustinBaird,MatthewMcDaniel,StevenWhite,NathanTatum,andLucasMarin. BioGears: A C++ library for whole body physiology simulations.Journal of Open Source Software, 5(56):2645, December 2020

work page 2020

[3] [3]

Ana Barragán-Montero, Umair Javaid, Gilmer Valdés, Dan Nguyen, Paul Desbor- des, Benoit Macq, Siri Willems, Liesbeth Vandewinckele, Mats Holmström, Fredrik Löfman, Steven Michiels, Kevin Souris, Edmond Sterpin, and John A. Lee. Arti- ficial intelligence and machine learning for medical imaging: A technology review. Physica Medica, 83:242–256, March 2021

work page 2021

[4] [4]

Hudson, Ehsan Adeli, Russ Altman, Simran Arora, Sydney von Arx, Michael S

Rishi Bommasani, Drew A. Hudson, Ehsan Adeli, Russ Altman, Simran Arora, Sydney von Arx, Michael S. Bernstein, Jeannette Bohg, Antoine Bosselut, Emma Brunskill, Erik Brynjolfsson, Shyamal Buch, Dallas Card, Rodrigo Castellon, Ni- ladri Chatterji, Annie Chen, Kathleen Creel, Jared Quincy Davis, Dora Dem- szky, Chris Donahue, Moussa Doumbouya, Esin Durmus, ...

arXiv 2022

[5] [5]

Sparks of Artificial General Intelligence: Early experiments with GPT-4, April 2023

Sébastien Bubeck, Varun Chandrasekaran, Ronen Eldan, Johannes Gehrke, Eric Horvitz, Ece Kamar, Peter Lee, Yin Tat Lee, Yuanzhi Li, Scott Lundberg, Harsha Nori, Hamid Palangi, Marco Tulio Ribeiro, and Yi Zhang. Sparks of Artificial General Intelligence: Early experiments with GPT-4, April 2023. arXiv:2303.12712 [cs]

arXiv 2023

[6] [6]

Theis, Alan Moses, and Bo Wang

MicaelaE.Consens,CameronDufault,MichaelWainberg,DuncanForster,Mehran Karimzadeh, Hani Goodarzi, Fabian J. Theis, Alan Moses, and Bo Wang. To Transformers and Beyond: Large Language Models for the Genome, November

work page

[7] [7]

The MIT Press, Cambridge, Mass, November 2016

Ian Goodfellow, Yoshua Bengio, and Aaron Courville.Deep Learning. The MIT Press, Cambridge, Mass, November 2016

work page 2016

[8] [8]

MOMENT: A Family of Open Time-series Foundation Models, February 2024

Mononito Goswami, Konrad Szafer, Arjun Choudhry, Yifu Cai, Shuo Li, and Ar- tur Dubrawski. MOMENT: A Family of Open Time-series Foundation Models, February 2024. arXiv:2402.03885 [cs]

arXiv 2024