❌

Normal view

Received — 15 September 2026 ⏭

Explaining Time Series Forecasting with Horizon-Resolved Attribution

✇cs.AI, q-bio.NC updates on arXiv.org
By: Seunghan Lee · Jun Seo · Jaehoon Lee · Junhyeok Kang · Sangjun Han · Sungdong Yoo · Minjae Kim · Tae Yoon Lim · Dongwan Kang · Hwanil Choi · Soonyoung Lee · Wonbin Ahn
14 September 2026 at 12:00
arXiv:2609.12639v1 Announce Type: cross Abstract: Recent advances in explaining time series (TS) models have produced methods that identify which past values a prediction depends on. However, most existing methods return a single importance vector, assuming that every predicted step depends on the same past values. In this paper, we show that this assumption does not hold, as different forecast steps depend on different past values. Motivated by this observation, we propose Horizon-Resolved eXplanation (HRX), which adds a horizon axis to the explanation, so that every forecast step receives its own importance map. HRX is a simple yet effective plug-in framework with three components: 1) an estimator that reads these maps out of any differentiable forecaster without modifying the TS backbone, 2) an evaluation protocol that validates the horizon axis by measuring how much a single forecast step changes when the inputs an importance map ranks highest are removed, and 3) a rank criterion that predicts in advance whether the axis is worth resolving on a given TS. We further show that this step-wise dependence is low-dimensional, as the explanations of all steps are built from a few shared maps whose number does not grow with the forecast length. Extensive experiments across various backbones and datasets show that the improvement comes from the horizon axis and holds for estimators of previous explanation methods. Code is available at https://github.com/seunghan96/HRX.
Received — 10 September 2026 ⏭

Divergent lipid utilization strategies of SARS-CoV-2 and MERS-CoV revealed by comparative multi-omics profiling of infected mouse lung tissues

✇Omics In Lung
By: Yeonseo Jang · Hyeran Kim · Yufei Li · Jae-Seung Lee · Jaehoon Lee · Han-Woong Lee · Nam-Hyuk Cho · Joo-Youn Cho
9 September 2026 at 18:00

Front Immunol. 2026 Aug 25;17:1902981. doi: 10.3389/fimmu.2026.1902981. eCollection 2026.

ABSTRACT

BACKGROUND: Coronaviruses (CoVs), including severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) and Middle East respiratory syndrome (MERS-CoV), cause respiratory infections with distinct clinical outcomes and case fatality rates. However, the molecular basis of these differences remains unclear. In this study, we sought to define virus-specific host metabolic programs by directly comparing multiomics profiles of the lungs of lethally infected mouse models.

METHODS: We performed integrated multiomics analyses, including untargeted metabolomics, transcriptomics, and targeted lipidomics, of lung tissues from human angiotensin-converting enzyme 2 (hiACE2)-human dipeptidyl peptidase 4 (hDPP4) double-knock-in (DKI) mice infected in SARS-CoV-2 or MERS-CoV. Data Integration Analysis and Biomarker discovery using Latent cOmponents (DIABLO) was applied across all three omics layers to identify key distinguishing molecular patterns. Additionally, in vitro lipid droplet kinetics were examined in infected Vero E6 cells to validate temporal differences in lipid remodeling.

RESULTS: We identified two distinct strategies for lipid utilization. SARS-CoV-2 infection showed strong activation of energy and amino acid metabolism at an early stage of infection (3 days post infection, DPI), whereas MERS-CoV infection was characterized by sustained alterations in lipid and nucleotide metabolism. Integrative DIABLO analysis of all three omics layers revealed that the key distinguishing features clustered into virus-specific molecular signatures: a triacylglycerol-lipid droplet-interferon axis for SARS-CoV-2 and a phospholipid-sphingolipid-membrane hub for MERS-CoV. In vitro lipid droplet kinetics in infected Vero E6 cells confirmed this temporal difference, with SARS-CoV-2 peaking earlier than MERS-CoV.

CONCLUSION: These findings show that β-CoVs exploit host lipid metabolism through virus-specific and time-dependent remodeling programs, providing a framework for understanding differential pathogenesis and developing host-directed antiviral strategies.

PMID:42712680 | PMC:PMC13550176 | DOI:10.3389/fimmu.2026.1902981

Divergent lipid utilization strategies of SARS-CoV-2 and MERS-CoV revealed by comparative multi-omics profiling of infected mouse lung tissues

✇(Multiomics OR Omics) AND (Lung OR gastric OR Hepatocellular)
By: Yeonseo Jang · Hyeran Kim · Yufei Li · Jae-Seung Lee · Jaehoon Lee · Han-Woong Lee · Nam-Hyuk Cho · Joo-Youn Cho
9 September 2026 at 18:00

Front Immunol. 2026 Aug 25;17:1902981. doi: 10.3389/fimmu.2026.1902981. eCollection 2026.

ABSTRACT

BACKGROUND: Coronaviruses (CoVs), including severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) and Middle East respiratory syndrome (MERS-CoV), cause respiratory infections with distinct clinical outcomes and case fatality rates. However, the molecular basis of these differences remains unclear. In this study, we sought to define virus-specific host metabolic programs by directly comparing multiomics profiles of the lungs of lethally infected mouse models.

METHODS: We performed integrated multiomics analyses, including untargeted metabolomics, transcriptomics, and targeted lipidomics, of lung tissues from human angiotensin-converting enzyme 2 (hiACE2)-human dipeptidyl peptidase 4 (hDPP4) double-knock-in (DKI) mice infected in SARS-CoV-2 or MERS-CoV. Data Integration Analysis and Biomarker discovery using Latent cOmponents (DIABLO) was applied across all three omics layers to identify key distinguishing molecular patterns. Additionally, in vitro lipid droplet kinetics were examined in infected Vero E6 cells to validate temporal differences in lipid remodeling.

RESULTS: We identified two distinct strategies for lipid utilization. SARS-CoV-2 infection showed strong activation of energy and amino acid metabolism at an early stage of infection (3 days post infection, DPI), whereas MERS-CoV infection was characterized by sustained alterations in lipid and nucleotide metabolism. Integrative DIABLO analysis of all three omics layers revealed that the key distinguishing features clustered into virus-specific molecular signatures: a triacylglycerol-lipid droplet-interferon axis for SARS-CoV-2 and a phospholipid-sphingolipid-membrane hub for MERS-CoV. In vitro lipid droplet kinetics in infected Vero E6 cells confirmed this temporal difference, with SARS-CoV-2 peaking earlier than MERS-CoV.

CONCLUSION: These findings show that β-CoVs exploit host lipid metabolism through virus-specific and time-dependent remodeling programs, providing a framework for understanding differential pathogenesis and developing host-directed antiviral strategies.

PMID:42712680 | PMC:PMC13550176 | DOI:10.3389/fimmu.2026.1902981

Received — 27 May 2026 ⏭

FinSTaR: Towards Financial Reasoning with Time Series Reasoning Models

✇cs.AI, q-bio.NC updates on arXiv.org
By: Seunghan Lee · Jun Seo · Jaehoon Lee · Sungdong Yoo · Minjae Kim · Tae Yoon Lim · Dongwan Kang · Hwanil Choi · Soonyoung Lee · Wonbin Ahn
26 May 2026 at 12:00
arXiv:2605.03460v2 Announce Type: replace Abstract: Time series (TS) reasoning models (TSRMs) have shown promising capabilities in general domains, yet they consistently fail on financial domain, which exhibit unique characteristics. We propose a general 2x2 capability taxonomy for TSRMs by crossing 1) single-entity vs. multi-entity analysis with 2) assessment of the current state vs. prediction of future behavior. We instantiate this taxonomy in the financial domain -- where the distinction between deterministic assessment and stochastic prediction is particularly critical -- as ten financial reasoning tasks, forming the FinTSR-Bench benchmark based on S&P stocks. To this end, we propose FinSTaR (Financial Time Series Thinking and Reasoning), trained on FinTSR-Bench with distinct chain-of-thought (CoT) strategies tailored to each category. For assessment, which is deterministic (i.e., computable from observable data), we employ Compute-in-CoT, a programmatic CoT that enables models to derive answers directly from raw prices. For prediction, which is inherently stochastic (i.e., subject to unobservable factors), we adopt Scenario-Aware CoT, which generates diverse scenarios before making a judgment, mirroring how financial analysts reason under uncertainty. The proposed method achieves 78.9% average accuracy on FinTSR-Bench, substantially outperforming LLM and TSRM baselines. Furthermore, we show that the four capability categories are complementary and mutually reinforcing through joint training, and that Scenario-Aware CoT consistently improves prediction accuracy over standard CoT. Code is publicly available at: https://github.com/seunghan96/FinSTaR.
Received — 26 March 2026 ⏭

Rethinking Multimodal Fusion for Time Series: Auxiliary Modalities Need Constrained Fusion

✇cs.AI, q-bio.NC updates on arXiv.org
By: Seunghan Lee · Jun Seo · Jaehoon Lee · Sungdong Yoo · Minjae Kim · Tae Yoon Lim · Dongwan Kang · Hwanil Choi · SoonYoung Lee · Wonbin Ahn
25 March 2026 at 12:00
arXiv:2603.22372v1 Announce Type: cross Abstract: Recent advances in multimodal learning have motivated the integration of auxiliary modalities such as text or vision into time series (TS) forecasting. However, most existing methods provide limited gains, often improving performance only in specific datasets or relying on architecture-specific designs that limit generalization. In this paper, we show that multimodal models with naive fusion strategies (e.g., simple addition or concatenation) often underperform unimodal TS models, which we attribute to the uncontrolled integration of auxiliary modalities which may introduce irrelevant information. Motivated by this observation, we explore various constrained fusion methods designed to control such integration and find that they consistently outperform naive fusion methods. Furthermore, we propose Controlled Fusion Adapter (CFA), a simple plug-in method that enables controlled cross-modal interactions without modifying the TS backbone, integrating only relevant textual information aligned with TS dynamics. CFA employs low-rank adapters to filter irrelevant textual information before fusing it into temporal representations. We conduct over 20K experiments across various datasets and TS/text models, demonstrating the effectiveness of the constrained fusion methods including CFA. Code is publicly available at: https://github.com/seunghan96/cfa/.
There are no more articles
❌