Baseline-Relative Counterfactual Refinement for Bit-Aware Visual Token Communication

Jia Guo; Xiaohan Zhao; Changwang Liu; Shuqing He; Chenyang Zhang; Bingchuan Zhao; Jinqi Zhu

cs.AIAug 17, 2026

Baseline-Relative Counterfactual Refinement for Bit-Aware Visual Token Communication

Authors: Jia Guo, Xiaohan Zhao, Changwang Liu, Shuqing He, Chenyang Zhang, Bingchuan Zhao, Jinqi Zhu

Organizations: School of Computer and Information Engineering, Tianjin Normal University, Tianjin, China · School of Information Science and Engineering, Linyi University, Linyi 276000, China

Abstract

Generative visual-token communication reduces transmission load by sending only selected discrete tokens and reconstructing missing content at the receiver. However, existing token-selection criteria based on local uncertainty, importance, or diversity do not directly determine whether changing the current selection improves the final reconstruction under the same packet budget. To address this problem, we propose Gated Counterfactual Refinement for Communication (GCR-C), a rollout-style correction layer over Local-MDL. GCR-C constructs a compact diversified candidate set, evaluates each candidate through matched full-budget Local-MDL continuation, and replaces the baseline action only when a positive baseline-relative reconstruction gain is obtained. Experiments on CIFAR-10, STL-10, a coded 5G-LDPC link, and a limited high-resolution Kodak transfer show that GCR-C consistently improves reconstruction quality at active low- and medium-rate operating points without increasing the realized packet rate, while remaining effective across changes in dataset, channel condition, resolution, token grid, and tokenizer. The results also reveal a clear quality--computation tradeoff due to the additional encoder-side counterfactual evaluation.

Figures & tables

Fig. 1: System architecture of GCR-C. A compact proposal is evaluated by full-budget Local-MDL completion, and an alternative action is accepted only when its baseline-relative advantage is positive. Selected tokens are packetized and transmitted through the coded link; receiver reconstruction is used only for offline evaluation.

Fig. 2: Validation diagnostics: headroom CDF (a), state headroom/P4 regret with image-cluster CIs (b), horizon quality–latency frontier (c), and proposal-source ablation (d).

Fig. 3: Proposal-size quality–compute frontier;

K=8

is the final compact operating point.

Nominal bpp	Method	Packet bpp	Tokens	PSNR (dB)	SSIM	$\Delta$ PSNR	95% CI	Win rate
0.20	Random	0.1986	11.41	13.215 $\pm$ 3.511	.382	+0.798	[+.552,+1.042]	.548
	Entropy-greedy	0.1986	12.37	13.319 $\pm$ 3.520	.392	+0.902	[+.674,+1.134]	.576
	Coverage-greedy	0.1992	11.00	12.889 $\pm$ 3.653	.365	+0.472	[+.237,+.706]	.506
	Adapted max–min diversity [ 21 ]	0.1984	12.49	11.903 $\pm$ 4.006	.333	$-$ 0.514	[ $-$ 0.775, $-$ 0.245]	.368
	Local-MDL	0.1985	12.02	12.417 $\pm$ 3.864	.395	0	—	—
	GCR-C	0.1987	12.01	13.920 $\pm$ 3.637	.436	+1.503	[+1.341,+1.670]	.974

TABLE I: Held-out CIFAR-10 results on 500 images. Values are mean

\pm

standard deviation for PSNR and mean SSIM. Random, Entropy-greedy, Coverage-greedy, and Adapted max–min diversity are matched baselines evaluated with the same packet protocol; the difference, CI, and win rate are paired against Local-MDL. At 0.44 bpp, GCR-C is identical to Local-MDL by policy; therefore, the win rate is 0 and the tie rate is 1.00.

Fig. 4: Held-out CIFAR-10 rate–distortion and paired GCR-C–Local-MDL 95% CIs; matched baselines are in Table I .

Rate	Channel	$E_{b}/N_{0}$ (H/M/E)	$\Delta$ PSNR (H/M/E)
0.20	AWGN	3.0/3.5/4.5	+.946/+1.194/+1.416
0.20	Block-Rayleigh	4.0/8.0/12.5	+.697/+1.071/+1.304
0.32	AWGN	2.5/3.0/3.5	+.378/+.539/+.647
0.32	Block-Rayleigh	4.0/7.5/12.0	+.394/+.450/+.630

TABLE II: Compact frozen-selection coded-link summary on the held-out CIFAR-10 split.

E_{b}/N_{0}

and

\Delta

PSNR are listed as hard/medium/easy triples; all 12 paired 95% CIs are positive. Complete FER, PSNR, and CI records are provided in Supplementary Table S4.

Point	Method	Cap (bits)	Packet bpp	Tokens	PSNR/ SSIM	$\Delta$ PSNR	95% CI	Win
Low	Local-MDL	350	.037907	24.912	11.998/.261	–	–	–
Low	GCR-C	350	.037911	24.754	12.620/.275	+.621	[.559,.687]	99.6%
Mid	Local-MDL	650	.070346	65.678	14.533/.358	–	–	–
Mid	GCR-C	650	.070350	65.630	14.738/.364	+.204	[.169,.243]	94.2%

TABLE III: STL-10 replication on a stratified 500-image held-out subset. Packet bpp is recomputed from complete packet fields;

\Delta

PSNR, CI, and win rate are paired GCR-C–Local-MDL differences.

Point	Method	Tok.	Pkt bpp	PSNR/SSIM	$\Delta$ PSNR [CI]
HR-Low	Local-MDL	87.708	.014204	9.349/.241	–
	GCR-C	87.667	.014201	10.382/.273	+1.033 [.641,1.456]
HR-Mid	Local-MDL	173.500	.025748	10.329/.238	–
	GCR-C	173.542	.025746	11.177/.290	+.847 [.369,1.417]

TABLE IV: Kodak-24 high-resolution transfer. Values are means over 24 images; Packet bpp is realized packet rate and the

\Delta

PSNR interval is a paired 10,000-resample image-level bootstrap CI.

Explore similar work

Selective Amortization of Full-Budget Counterfactual Reasoning for Visual Token Communication
Sep 25, 2026Qinglei Qi, Zhihe Liang, Fengzhan Jing +5Generative Image Compression Adaptive Computation
School of Artificial Intelligence, Nanyang Normal University, Nanyang, China · School of Computer and Information Engineering, Tianjin Normal University, Tianjin, China · School of Information Science and Engineering, Linyi University, Linyi 276000, China
Beyond Token Accuracy: Prioritizing What Matters for Visual Reconstruction
Oct 2, 2026Zhicheng Liu, Zhouxiang Zhao, Chenliang Wu +2Image Reconstruction Wireless Resource Allocation
Zhejiang University
Token Encoding for Semantic Recovery
Apr 14, 2026Jingzhi Hu, Ouya Wang, Geoffrey Ye LiSemantic Communication Wireless Communications
Department of Electrical and Electronic Engineering, Imperial College London, London SW7 2AZ, UK.

Sep 25, 2026J/K move · Enter open · S save

Selective Amortization of Full-Budget Counterfactual Reasoning for Visual Token Communication

Qinglei Qi, Zhihe Liang, Fengzhan Jing, Shenao Zhu+4

School of Artificial Intelligence, Nanyang Normal University, Nanyang, China · School of Computer and Information Engineering, Tianjin Normal University, Tianjin, China · School of Information Science and Engineering, Linyi University, Linyi 276000, China

Generative Image Compression Adaptive Computation Visual Token Compression Counterfactual Evaluation

Open paper AskPDF arXiv

Generative image communication transmits compact semantic tokens under a limited packet budget, where token selection directly affects the final reconstruction quality after the complete packet is decoded. However, accurately estimating the terminal value of every candidate token requires repeated receiver-side reconstruction, resulting in substantial encoder-side computation. To address this problem, we propose ACV-Gate, an adaptive candidate evaluation framework that learns to approximate full-budget counterfactual evaluation and selectively assigns exact evaluations to the most informative candidates. Specifically, a set-aware student is trained using terminal advantages and regrets to predict candidate rankings directly, while a selective refinement mechanism evaluates only a bounded candidate set containing both Local-MDL and direct actions; cost-based thresholds further enable explicit control of the average evaluation workload. Experiments on CIFAR-10 show that ACV-Gate consistently improves reconstruction quality while substantially reducing candidate evaluations; at 0.20 bpp, the primary adaptive configuration improves PSNR over LocalMDL by 0.636 dB with only 2.13 candidate evaluations per image, corresponding to 27.60% of the calls required by the Exact-Full expert. Matched-candidate comparisons, synchronized GPU measurements, and evaluations on STL-10 and 384 *384 scale transfer further demonstrate consistent quality computation trade-offs, with particularly pronounced gains at low bit rates. These results show that combining terminal-value learning with selective candidate evaluation provides an effective and controllable mechanism for allocating encoder computation in packet-constrained generative image communication.

Abstract

Explore similar work

Selective Amortization of Full-Budget Counterfactual Reasoning for Visual Token Communication

Beyond Token Accuracy: Prioritizing What Matters for Visual Reconstruction

Token Encoding for Semantic Recovery