icassp2027-experiment/manuscript/main/main.tex

% ICASSP 2027 content skeleton. The final paper must use the official 2027
% author kit and contain no visible \todo{} markers. Planned experiments are
% intentionally marked rather than presented as completed work.
\documentclass{article}
\usepackage{spconf}
\usepackage{amsmath,amssymb,graphicx,hyperref,xcolor}

\newcommand{\todo}[1]{\textcolor{red}{\textbf{[TODO: #1]}}}
\newcommand{\vect}[1]{\boldsymbol{#1}}
\setlength{\emergencystretch}{2em}

\title{Is It My Turn? Causal Local Evidence Adaptation of Omni Models for
Egocentric Floor Forecasting}

% TODO: Insert the final author list and affiliations after the ICASSP 2027
% review policy and author order are finalized.
\name{Internal Manuscript Draft}
\address{}

\begin{document}
\ninept
\maketitle

\begin{abstract}
Real-time audiovisual agents must decide not only what to say, but whether an
emerging floor opportunity belongs to them. This is difficult in multiparty
egocentric interaction: the target wearer is behind the camera, while the
visible people provide only indirect evidence about whether the surrounding
floor remains occupied. We study strictly causal forecasting of the wearer's
future speech activity every 200~ms. Although a general-purpose Omni model
provides rich long-context audio--visual representations, its global readout is
not explicitly trained to expose sub-second, wearer-relative floor evidence.
We therefore study a candidate causal local floor adapter that combines the global
Omni state with short-window mixed audio and an anonymous set of visible-person
observations, without assuming persistent speaker identity. On EasyCom,
\todo{report the final three-seed dense and competing-risk-censored event
results: false-entry reduction at matched wearer-onset recall, onset latency,
yield behavior, confidence intervals, and oracle-gap recovery}. Controlled
\todo{time-shift/remove-vision/same-rate or capacity-matched} interventions
show that \todo{state the supported mechanism, without claiming addressee or
speaker identity}. These results demonstrate \todo{the narrow conclusion
actually supported by the final experiments}.
\end{abstract}

\begin{keywords}
conversational agents, turn-taking, egocentric video, multimodal adaptation,
voice activity projection
\end{keywords}

\section{Introduction}
\label{sec:intro}

Natural turn-taking requires a conversational system to anticipate not only
when speech will stop, but whether a floor opportunity actually belongs to it.
In a dyadic exchange, a pause after one interlocutor finishes often invites the
other to respond. In a group conversation, the same pause may instead mark a
handover between two other participants. An egocentric agent that reacts only
to silence or aggregate speech may therefore interrupt a conversation that
should continue without it. We study \emph{egocentric floor forecasting}:
continuously predicting the near-future speech participation of the
first-person wearer from strictly causal audio--visual observations.

Voice Activity Projection (VAP) introduced continuous prediction of future
voice activity for turn-taking \cite{ekstedt2022vap}. EgoSpeak brought this
formulation to causal, untrimmed first-person streams, using future camera-
wearer activity as an imitation target \cite{kim2025egospeak}. Complementary
work adds visual gaze, head pose, and facial behavior to predictive turn-taking
\cite{russell2025mmvap}, while MuVAP combines global acoustic dynamics with
visible face tracks to forecast visible floor holders in multiparty scenes
\cite{qi2026muvap}. Our target has a different role structure: the participant
whose future action is predicted is the one participant absent from the
visible set. The visible people are contextual evidence for the unseen
wearer's decision, not candidate output identities.

This role asymmetry exposes an adaptation problem for general-purpose Omni
models. Omni pretraining favors globally integrated representations useful for
multimodal understanding and generation. Egocentric floor control instead
depends on timely prosodic and visible-person activity at sub-second scale,
including whether other participants continue to occupy the floor. Increasing
model capacity or applying a generic parameter-efficient update does not
explicitly preserve or route this evidence to a dense wearer-relative
prediction head. The open question is therefore not merely whether an Omni
backbone can be applied to EgoSpeak, but what task structure is still required
to turn general audiovisual understanding into reliable floor control.

Our candidate decomposition separates long-context understanding from
short-window floor evidence. A
causal global Omni state represents the broader scene and conversation. In
parallel, a lightweight local path encodes recent mixed audio and a changing,
anonymous set of visible-person observations. A residual adapter makes this
local evidence accessible to the common ten-horizon activity head. The primary
variant deliberately avoids cross-time identity: the current development
oracle supports timely aggregate visible-other activity but not person
continuity. Track-aware structure is considered only if it defeats no-ID,
identity-corrupted, and capacity-matched controls.

The final paper will claim the following three contributions only if the
corresponding TODO gates pass:
\begin{itemize}
  \item We identify and test a temporal--role mismatch between a global Omni
  readout and causal egocentric floor forecasting for an invisible wearer.
  \item We evaluate a wearer-relative causal local floor adapter that combines
  long-context Omni information with short-window audio and anonymous
  visible-person evidence.
  \item We evaluate whether the adapter reduces false wearer entries during
  multi-other exchanges at matched valid-onset recall, using causal alignment
  interventions and competing-risk-censored entry/yield analysis.
\end{itemize}

\begin{figure}[t]
\centering
\fbox{\parbox[c][1.00in][c]{0.94\columnwidth}{\centering
\todo{Draw only the selected method: causal AV to the long-context Omni path;
the short-window mixed-audio and anonymous person-set path; residual fusion;
the common $10\times3$ head; and the wearer-silent/active event interpretation.
Mark physical cutoffs, trainable modules, person masks, and NULL/OFFSCREEN.
Do not draw oracle labels, rejected track variants, or multiple backbones.}}}
\caption{Long-context Omni understanding and short-window local floor evidence
for wearer-relative future-activity prediction.}
\label{fig:method}
\end{figure}

\section{Method}
\label{sec:method}

\subsection{Causal wearer-relative projection}

At decision time $t$, the model observes only the stream
$\mathcal X_{\leq t}$ and predicts $K=10$ future bins at $\Delta=0.2$~s. For
bin $k$, the multilabel target
$\vect y_{t,k}=[y^{\mathrm{bg}},y^{\mathrm{w}},y^{\mathrm{o}}]$ denotes
background, camera-wearer activity, and activity by one or more other people.
Wearer--other overlap is allowed. The model outputs
\begin{equation}
 \hat{\vect Y}_t=f_\theta(\mathcal X_{\leq t})
 \in[0,1]^{K\times3}.
 \label{eq:task}
\end{equation}
The same wearer-activity forecast has two controller interpretations. In a
wearer-silent opportunity, $\hat y^{\mathrm w}_{t,1}$ is an entry score. While
the wearer is active, $1-\hat y^{\mathrm w}_{t,1}$ is a near-term yield score.
The current self-state is used to define these observable event strata, not as
an unreported oracle model input. \todo{If the final adapter explicitly uses
self-state, define its strictly causal and deployment-available source and add
a matched no-state ablation; otherwise keep self-state evaluation-only.}

\subsection{Global Omni baseline}

Let $\vect g_t\in\mathbb R^{d_g}$ be the causal global readout of
Qwen2.5-Omni at the physical cutoff $t$ \cite{xu2025qwenomni}. The frozen
baseline trains only a dimension-matched prediction head. A matched PEFT/LoRA
baseline updates a pre-registered subset of the Omni model under the same
examples, context, optimizer budget, head, and evaluator. Its purpose is to
test whether generic parameter adaptation is sufficient; one negative recipe
is not interpreted as a general limitation of LoRA or all Omni models.

\todo{Freeze the exact Qwen2.5-Omni-3B revision, processor, input rates,
streaming/cache semantics, readout layer, context, dtype, LoRA targets/rank,
trainable parameter budget, and cutoff audit. Add Qwen2.5-Omni-7B only as a
capacity-scaling control. Do not add another backbone family unless it is
needed to support and experimentally earns a cross-family claim.}

\subsection{Causal local floor adapter}

The candidate local path operates on the most recent $L_\ell$ seconds. At local instant
$\tau\leq t$, a causal audio encoder produces $\vect a_\tau$ from mixed audio.
For the anonymous visible-person set
$\mathcal P_\tau=\{(\vect x_{\tau,i},\vect b_{\tau,i})\}_{i=1}^{N_\tau}$,
a shared visual encoder and masked set operator produce
\begin{equation}
 \vect v_\tau=\mathrm{SetPool}
 \left(\{E_v(\vect x_{\tau,i},\vect b_{\tau,i})\}_{i=1}^{N_\tau}
 \cup\{\vect v_\varnothing\}\right),
 \label{eq:set}
\end{equation}
where $\vect v_\varnothing$ handles an empty or offscreen visible set. The
operator is permutation invariant and uses no participant-ID embedding or
cross-time identity. A causal short-window encoder summarizes local dynamics,
\begin{equation}
 \vect \ell_t=F_\ell
 \left(\{[\vect a_\tau;\vect v_\tau]\}_{\tau\in(t-L_\ell,t]}\right).
 \label{eq:local}
\end{equation}
The adapter preserves the global initialization through a zero-initialized
residual output,
\begin{equation}
 \vect z_t=\vect g_t+W_o\,
 \sigma\!\left(W_g\vect g_t+W_\ell\vect\ell_t\right),
 \qquad W_o\vert_{\mathrm{init}}=0,
 \label{eq:fusion}
\end{equation}
and the common head maps $\vect z_t$ to Eq.~\eqref{eq:task}. This implements an
evidence-access hypothesis rather than adding an unconstrained second model:
the global path retains broad context, while the local path preserves timely
floor evidence.

\todo{Freeze $L_\ell$, local audio/video rates, crop source and resolution,
set pooling, causal temporal encoder, masks, reset rules, parameter count, and
latency. If the local path runs faster than the 5-Hz global/output grid, compare
it with a same-parameter 5-Hz variant before using the term ``dual-rate.'' Use
provided causal boxes only as a clearly labeled method-isolation condition;
online detection/tracking is not a paper contribution unless actually run.}

\subsection{Training}

The primary objective is class-balanced binary cross entropy over valid
horizons and activity channels,
\begin{equation}
 \mathcal L_{\mathrm{act}}=
 \sum_{t,k,c}m_{t,k}w_c\,
 \mathrm{BCE}(\hat y^c_{t,k},y^c_{t,k}).
 \label{eq:loss}
\end{equation}
No transcript, target-of-speech label, close microphone, future frame, or
per-person speech label enters the deployable model. \todo{Add an active-speaker
or event auxiliary loss only if it improves the frozen primary protocol and
its label/input boundary is fully reported.}

\section{Experiments} 
\label{sec:experiments} 

\subsection{Dataset and Data Preparation}

We evaluate on EasyCom following the EgoSpeak split, using Sessions 4--12 for training and development and Sessions 1--3 for final evaluation \cite{donley2021easycom}. All hyperparameters, event windows, operating thresholds, and checkpoints are selected without access to the test sessions. The input stream is converted to strictly causal examples at 5~Hz. Each decision point is paired with ten future 0.2-s activity bins for background, camera wearer, and aggregate non-wearer speech. Provided person boxes are used only to construct the visible-person inputs; participant identity, close-mic audio, transcripts, and future observations are not available to the model. \todo{Report dataset duration, number of training/test examples and events, and the frozen development/calibration split.}

\subsection{Model Configuration and Training}

Our main backbone is Qwen2.5-Omni-3B. We compare a frozen global Omni readout, matched PEFT/LoRA adaptation, global features with recent local audio, a parameter- and history-matched temporal local branch, and the proposed causal local audio--visual adapter. Qwen2.5-Omni-7B is used only as a capacity-scaling control. All learned variants use the same causal context, prediction head, training examples, and class-balanced binary cross-entropy objective in Eq.~\eqref{eq:loss}. Unless otherwise stated, model selection is performed on the development data and final numbers are averaged over three random seeds. \todo{Freeze context length, local window/rates, optimizer, epochs, LoRA configuration, trainable parameter counts, and inference latency.}

\subsection{Evaluation Metrics}

We evaluate both dense future-activity prediction and event-level floor
control. For comparability with EgoSpeak, we report per-frame AP at
$0.2$--$2.0$~s horizons and their mean mAP.

For event-level evaluation, a \emph{wearer entry} is a transition from wearer
silence to speech, while a \emph{wearer yield} is a transition from speech to
silence. An entry prediction is considered correct if it is committed within
the predefined response window of a ground-truth wearer onset. We report entry
recall, false entries per minute of eligible wearer-silent time, and signed
onset latency. Yield events are evaluated analogously using recall and latency.

Multiparty interaction introduces a competing-risk condition. After another
speaker releases the floor, the current wearer-entry opportunity is terminated
if another non-wearer participant starts speaking before the wearer does.
Predictions after this competing onset are therefore not matched to the
previous opportunity. This prevents an other-to-other handover from being
treated as an open floor for the wearer.

Operating thresholds are selected on the development split and frozen for test
evaluation. Our primary behavioral metric is false wearer entries per minute at
a matched wearer-entry recall of $0.8$; dense mAP and wearer-yield performance
are reported as complementary measures.


\section{Results and Analysis}
\label{sec:results}

\subsection{Main Comparison}

Table~\ref{tab:main} compares the proposed causal local audiovisual adapter
with an EgoSpeak reproduction, the frozen Qwen2.5-Omni-3B global state $G$,
matched PEFT, $G$ plus recent local audio, and a parameter- and history-matched
temporal local branch. Qwen2.5-Omni-7B is included only as a capacity-scaling
control. All methods use the same split, causal inputs, targets, and evaluation
protocol. Event thresholds are selected on training-side calibration data and
frozen before test evaluation. \todo{Freeze all baseline implementations and
verify identical targets, causal validity, and evaluation.}

\begin{table}[t]
\caption{Confirmatory EasyCom results. Higher mAP and yield recall are better;
lower false entries and onset latency are better. FP/min is measured at
matched valid-onset recall $R=0.8$.}
\label{tab:main}
\centering
\small
\resizebox{\columnwidth}{!}{%
\begin{tabular}{lccccc}
\hline
Method & 0.2-s mAP $\uparrow$ & Mean mAP $\uparrow$ & FP/min $\downarrow$
& Onset lat. $\downarrow$ & Yield R $\uparrow$\\
\hline
EgoSpeak reproduction       & \todo{--} & \todo{--} & \todo{--} & \todo{--} & \todo{--}\\
Omni-3B frozen $G$          & \todo{--} & \todo{--} & \todo{--} & \todo{--} & \todo{--}\\
Omni-3B matched PEFT        & \todo{--} & \todo{--} & \todo{--} & \todo{--} & \todo{--}\\
$G+$local audio             & \todo{--} & \todo{--} & \todo{--} & \todo{--} & \todo{--}\\
$G+$matched temporal local  & \todo{--} & \todo{--} & \todo{--} & \todo{--} & \todo{--}\\
$G+$causal local AV (ours)  & \todo{--} & \todo{--} & \todo{--} & \todo{--} & \todo{--}\\
Omni-7B scaling control     & \todo{--} & \todo{--} & \todo{--} & \todo{--} & \todo{--}\\
\hline
\end{tabular}}
\end{table}

\textbf{Dense forecasting.}
The proposed local AV adapter improves 0.2-s mAP from
\todo{$X$ to $Y$} over the strongest matched baseline and improves
mean mAP by \todo{$Z$ points}.
The gain is largest at \todo{short prediction horizons / transition-adjacent
frames / specify actual observation}, indicating that local audiovisual
evidence primarily improves \todo{the interpretation supported by the data}.

\textbf{Floor-entry behavior.}
At matched valid-onset recall $R=0.8$, the proposed adapter reduces
multi-other false entries from \todo{$X$} to \todo{$Y$} FP/min
(\todo{$Z\%$ relative reduction}).
This reduction is obtained while \todo{maintaining / improving} onset latency
and yield recall.
Thus, the improvement cannot be explained solely by a more conservative
decision threshold.

\textbf{Comparison with adaptation baselines.}
Matched PEFT yields \todo{result}, whereas adding recent local audio alone
yields \todo{result}.
The parameter- and history-matched temporal branch achieves
\todo{result}.
The remaining gap between this control and the proposed adapter therefore
suggests that \todo{correctly timed audiovisual local evidence contributes
beyond additional temporal capacity}.
\todo{If this difference is not statistically supported, state instead that
the local temporal branch accounts for most of the gain.}

\subsection{Information Opportunity}

Before designing the deployable adapter, we measured whether local floor
information was useful at all using an analysis-only oracle.
The oracle receives perfectly timed visible-other speech activity through a
non-deployable pathway and is never used in the main comparison.

On development Session~12, aggregate visible activity improved 0.2-s mAP by
6.97 points and ten-horizon mean mAP by 2.91 points over the matched
global-plus-local-audio baseline.
In contrast, visibility alone and one-second-delayed activity remained close
to the audio baseline, while person-specific histories did not outperform
identity-shuffled histories.

These controls indicate that the useful signal is primarily
\emph{timely local floor evidence}, rather than participant identity or
static visibility.
The proposed adapter is therefore designed to approximate this information
using only causally available audiovisual observations.

\todo{After completing all seeds, report mean $\pm$ confidence interval here.
If the oracle effect is unstable, describe it qualitatively and omit the exact
numbers from the four-page version.}

\subsection{Mechanism Analysis}

To determine whether the gain arises from correctly aligned visual floor
evidence rather than additional parameters or generic visual context, we
evaluate the controls in Table~\ref{tab:ablation}.

\begin{table}[t]
\caption{Mechanism controls for the selected Omni-3B local adapter. All
variants use the same prediction head and comparable temporal capacity.}
\label{tab:ablation}
\centering
\small
\resizebox{\columnwidth}{!}{%
\begin{tabular}{lccc}
\hline
Local condition & Mean mAP $\uparrow$ & FP/min@$R=.8$ $\downarrow$
& Yield R $\uparrow$\\
\hline
Correct causal local AV       & \todo{--} & \todo{--} & \todo{--}\\
Remove local vision           & \todo{--} & \todo{--} & \todo{--}\\
Time-shift local vision       & \todo{--} & \todo{--} & \todo{--}\\
Global/full-frame visual pool & \todo{--} & \todo{--} & \todo{--}\\
Same-rate/capacity control    & \todo{--} & \todo{--} & \todo{--}\\
\hline
\end{tabular}}
\end{table}

Removing local vision changes \todo{metric} by \todo{$X$}, while temporally
shifting the visual stream by \todo{$\Delta t$} changes it by \todo{$Y$}.
Replacing person-local visual evidence with a global/full-frame visual pool
produces \todo{result}.
Together, these comparisons \todo{support / do not support} the hypothesis
that the benefit depends specifically on correctly timed local visual
evidence.

The same-rate/capacity control achieves \todo{result}, showing that
\todo{the gain cannot / can largely} be attributed to additional temporal
parameters or higher-rate processing alone.

\todo{Report paired intervals and the supported interpretation. If correct
and shifted/removed vision do not separate, drop the audiovisual
evidence-routing claim. If dense mAP improves but event behavior does not, do
not claim improved floor control. If legitimate onset or yield degrades, do
not describe lower false entry as success.}

\subsection{Scaling Control}

Increasing the frozen Omni backbone from 3B to 7B yields
\todo{$X$ points} in 0.2-s mAP and \todo{$Y$} FP/min at matched recall.
The proposed 3B adapter \todo{remains better than / approaches / is surpassed
by} the 7B scaling control.

This result suggests that \todo{architectural access to local causal evidence
provides a complementary benefit to backbone scaling}.
Because both backbones belong to the Qwen2.5-Omni family, we do not claim that
this conclusion generalizes to all omni-modal models.

\subsection{Limitations and Qualitative Analysis}

The results support \todo{precise supported claim}, but several limitations
remain.
First, \todo{EasyCom size/session diversity limitation}.
Second, the adapter relies on \todo{state the actual visible-person box or
tracking assumption}.
Third, \todo{failure condition: occlusion, overlapping speakers, rapid floor
exchange, etc.} remains challenging.

\todo{Add one compact, pre-selected qualitative timeline only if space permits:
one multi-other false-entry improvement paired with one preserved real wearer
onset or one failure. Examples illustrate aggregate rates and never establish
the claim by themselves.}

\section{Conclusion}
\label{sec:conclusion}

This study asks whether a general-purpose Omni representation can be adapted
for causal floor forecasting of an invisible egocentric wearer. The candidate
method separates long-context Omni understanding from short-window local floor
evidence without assuming persistent speaker identity. \todo{State the final
learned-method improvement, matched-recall false-entry result, onset/yield
non-degradation, causal-control evidence, and main limitation. If no learned
variant passes these tests, reframe the submission as an information audit or
do not make the method claim.}

\bibliographystyle{IEEEbib}
\bibliography{references}

\end{document}