@inproceedings{niess-kern-2026-spandiffusion,
title = "{S}pan{D}iffusion: Flow Matching over Continuous Span Masks for Financial Causal Question Answering",
author = "Niess, Georg and
Kern, Roman",
editor = "El-Haj, Mo and
Moreno Sandoval, Antonio and
Garcia-Serrano, Ana and
Chen, Chung-Chi and
Rayson, Paul and
Torterolo Orta, Yanco Amor and
Martinez, Paloma and
Porta, Jordi",
booktitle = "The 7th Financial Narrative Processing Workshop",
month = may,
year = "2026",
address = "Palma de Mallorca, Spain",
publisher = "European Language Resources Association (ELRA)",
url = "https://aclanthology.org/2026.fnp-1.15/",
doi = "10.63317/4bm7v3ccgmjy",
pages = "146--151",
abstract = "We present SpanDiffusion, a continuous diffusion approach to extractive causal question answering for the FinCausal 2026 shared task. SpanDiffusion uses two Gaussian masks, continuous signals with peaks at the answer start and end positions, and learns to denoise them from pure noise through a dedicated transformer conditioned on frozen DeBERTa-v3-large embeddings with LoRA adapters (1.6M parameters). By replacing Denoising Diffusion Probabilistic Models (DDPM) with flow matching (rectified flow), we reduce denoising to only 20 Euler steps at inference. A systematic ablation across six diffusion variants and a span-classification baseline shows that LoRA adaptation is the dominant factor (+34 Exact Match points), followed by flow matching (+5.5 EM). However, the standard span classifier (85.8{\%} EM) outperforms our best diffusion model (83.0{\%} EM), suggesting that the denoiser does not yet justify its added complexity. We discuss tradeoffs between the interpretability of diffusion trajectories and classification accuracy."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="niess-kern-2026-spandiffusion">
<titleInfo>
<title>SpanDiffusion: Flow Matching over Continuous Span Masks for Financial Causal Question Answering</title>
</titleInfo>
<name type="personal">
<namePart type="given">Georg</namePart>
<namePart type="family">Niess</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Roman</namePart>
<namePart type="family">Kern</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-05</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>The 7th Financial Narrative Processing Workshop</title>
</titleInfo>
<name type="personal">
<namePart type="given">Mo</namePart>
<namePart type="family">El-Haj</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Antonio</namePart>
<namePart type="family">Moreno Sandoval</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ana</namePart>
<namePart type="family">Garcia-Serrano</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Chung-Chi</namePart>
<namePart type="family">Chen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Paul</namePart>
<namePart type="family">Rayson</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yanco</namePart>
<namePart type="given">Amor</namePart>
<namePart type="family">Torterolo Orta</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Paloma</namePart>
<namePart type="family">Martinez</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jordi</namePart>
<namePart type="family">Porta</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>European Language Resources Association (ELRA)</publisher>
<place>
<placeTerm type="text">Palma de Mallorca, Spain</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>We present SpanDiffusion, a continuous diffusion approach to extractive causal question answering for the FinCausal 2026 shared task. SpanDiffusion uses two Gaussian masks, continuous signals with peaks at the answer start and end positions, and learns to denoise them from pure noise through a dedicated transformer conditioned on frozen DeBERTa-v3-large embeddings with LoRA adapters (1.6M parameters). By replacing Denoising Diffusion Probabilistic Models (DDPM) with flow matching (rectified flow), we reduce denoising to only 20 Euler steps at inference. A systematic ablation across six diffusion variants and a span-classification baseline shows that LoRA adaptation is the dominant factor (+34 Exact Match points), followed by flow matching (+5.5 EM). However, the standard span classifier (85.8% EM) outperforms our best diffusion model (83.0% EM), suggesting that the denoiser does not yet justify its added complexity. We discuss tradeoffs between the interpretability of diffusion trajectories and classification accuracy.</abstract>
<identifier type="citekey">niess-kern-2026-spandiffusion</identifier>
<identifier type="doi">10.63317/4bm7v3ccgmjy</identifier>
<location>
<url>https://aclanthology.org/2026.fnp-1.15/</url>
</location>
<part>
<date>2026-05</date>
<extent unit="page">
<start>146</start>
<end>151</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T SpanDiffusion: Flow Matching over Continuous Span Masks for Financial Causal Question Answering
%A Niess, Georg
%A Kern, Roman
%Y El-Haj, Mo
%Y Moreno Sandoval, Antonio
%Y Garcia-Serrano, Ana
%Y Chen, Chung-Chi
%Y Rayson, Paul
%Y Torterolo Orta, Yanco Amor
%Y Martinez, Paloma
%Y Porta, Jordi
%S The 7th Financial Narrative Processing Workshop
%D 2026
%8 May
%I European Language Resources Association (ELRA)
%C Palma de Mallorca, Spain
%F niess-kern-2026-spandiffusion
%X We present SpanDiffusion, a continuous diffusion approach to extractive causal question answering for the FinCausal 2026 shared task. SpanDiffusion uses two Gaussian masks, continuous signals with peaks at the answer start and end positions, and learns to denoise them from pure noise through a dedicated transformer conditioned on frozen DeBERTa-v3-large embeddings with LoRA adapters (1.6M parameters). By replacing Denoising Diffusion Probabilistic Models (DDPM) with flow matching (rectified flow), we reduce denoising to only 20 Euler steps at inference. A systematic ablation across six diffusion variants and a span-classification baseline shows that LoRA adaptation is the dominant factor (+34 Exact Match points), followed by flow matching (+5.5 EM). However, the standard span classifier (85.8% EM) outperforms our best diffusion model (83.0% EM), suggesting that the denoiser does not yet justify its added complexity. We discuss tradeoffs between the interpretability of diffusion trajectories and classification accuracy.
%R 10.63317/4bm7v3ccgmjy
%U https://aclanthology.org/2026.fnp-1.15/
%U https://doi.org/10.63317/4bm7v3ccgmjy
%P 146-151
Markdown (Informal)
[SpanDiffusion: Flow Matching over Continuous Span Masks for Financial Causal Question Answering](https://aclanthology.org/2026.fnp-1.15/) (Niess & Kern, FNP 2026)
ACL