X-Tree: Tokenizing Reusable Experience for Efficient Agent Generalization

X-Tree learns reusable hierarchical skills from agent trajectories and improves success rates up to 5.8% across web and science benchmarks.

Sitao Cheng, Xunjian Yin, Zhiyuan Sun, Yuxuan Li and 3 more

Published Sep 26, 2026 · 0 citations · ▲ 76 on Hugging Face · Code ★ 2

Sample-Efficient Optimization over Generative Priors via Coarse Learnability

Pranjal Awasthi, Sreenivas Gollapudi, Ravi Kumar, Kamesh Munagala

Sydney Poster Session 1, Tue, Dec 8, 10:00 AM–1:00 PM, Hall 1-4 · Published 2026

Extending ROC Analysis to Uncertainty-Aware Risk Prediction with an Interval-Based AUC (iAUC)

Yuqi Li, Matthew Engelhard

Sydney Poster Session 2, Tue, Dec 8, 5:00 PM–8:00 PM, Hall 1-4 · Published 2026

FocusVLA: Hijacking Attention to Break Visual Token Pruning in Vision-Language-Action Models

Yanhui Li, Qianpu Sun, Qi Zhou, Chenru Jiang and 2 more

Sydney Poster Session 2, Tue, Dec 8, 5:00 PM–8:00 PM, Hall 1-4 · Published 2026

Learning Biological Hierarchies in Single-Cell Foundation Models

Xiangyu Guo, Ricardo Henao

Atlanta Poster Session 2, Wed, Dec 9, 4:30 PM–7:30 PM, Hall C1 · Published 2026

The Type Theory of Stationary MDPs: Rare Events and Uncertainty Quantification

Imon Banerjee, Sayak Chakrabarty, Ramkrishna Jyoti Samanta, Riddhiman Bhattacharya

Atlanta Poster Session 3, Thu, Dec 10, 10:00 AM–1:00 PM, Hall C1 · Published 2026

Co-Evolving Interpolants and Flows via Path-Flow Alignment

Zeyu M Li, William X Chen, Xiang Cheng

Atlanta Poster Session 3, Thu, Dec 10, 10:00 AM–1:00 PM, Hall C1 · Published 2026

Beyond Task Success: Probing Cognitive Primitives in Web Agents

Xunjian Yin, Tianchen Guan, Jinao Wang, Weili Cao and 7 more

Atlanta Poster Session 3, Thu, Dec 10, 10:00 AM–1:00 PM, Hall C1 · Published 2026

Diverse Representative Rashomon Sets for Sparse Generalized Additive Models

Varun Babbar, Christopher Li, Chudi Zhong, Cynthia Rudin

Sydney Poster Session 5, Thu, Dec 10, 10:00 AM–1:00 PM, Hall 1-4 · Published 2026

What the Geometry of Good Models Tells Us

Alexis Fox, Samuel Orellana Mateo, Krish Yadav, Yiyang Sun and 2 more

Sydney Poster Session 3, Wed, Dec 9, 10:00 AM–1:00 PM, Hall 1-4 · Published 2026

Efficient Algorithms For Fully Dynamic Bipartite Matching In Metric Spaces

Pankaj Agarwal, Oliver Chubet, Sharath Raghvendra, Arian Zamani

Atlanta Poster Session 2, Wed, Dec 9, 4:30 PM–7:30 PM, Hall C1 · Published 2026

Ordered Policy Optimization

Dongkyu (Derek) Cho, Pan Xu

Sydney Poster Session 2, Tue, Dec 8, 5:00 PM–8:00 PM, Hall 1-4 · Published 2026

MARS: Multi-resolution Adaptive Routing for Sequential Recommendation

Ming Yin, Sixun Dong, Yudong Liu, Wenyun Yang and 2 more

Sydney Poster Session 4, Wed, Dec 9, 5:00 PM–8:00 PM, Hall 1-4 · Published 2026

NPCBench: A Clinical Apprenticeship Benchmark for Guideline-Constrained Care-Pathway Reasoning in Nasopharyngeal Carcinoma

Pengkai Wang, Wei-Wei Zhang, Yan Li, Min Tang and 12 more

Sydney Poster Session 3, Wed, Dec 9, 10:00 AM–1:00 PM, Hall 1-4 · Published 2026

A$^2$IQL: Adaptive Asymmetric Implicit Q Learning for Automated Warehouse Consolidation

Guangyi Liu, Andrea Angiuli, Mirko Ristivojevic, Joseph W Durham and 2 more

Atlanta Poster Session 1, Wed, Dec 9, 10:00 AM–1:00 PM, Hall C1 · Published 2026

Multilevel and Sequential Monte Carlo for Training-Free Diffusion Guidance

A sequential Monte Carlo framework with multilevel variance reduction provides unbiased diffusion guidance, achieving state-of-the-art training-free conditional generation with lower cost.

Aidan Gleich, Scott C Schmidler

Paris Poster Session 1, Wed, Dec 9, 12:30 PM–2:30 PM, Paris Poster Hall · Published 2026

PRISM: Rethinking Atmospheric Scattering Reconstruction as a Unified Understanding and Restoration Model for Real-world Dehazing

PRISM proposes a unified physically structured framework for real-world dehazing via proximal scattering atmosphere reconstruction, online non-uniform haze synthesis, and selective self-distillation adaptation to achieve competitive restoration results.

Chengyu Fang, Chunming He, Yuelin Zhang, Chubin Chen and 5 more

Sydney Poster Session 2, Tue, Dec 8, 5:00 PM–8:00 PM, Hall 1-4 · Published 2026

AMPS: Adaptive Modality Preference Steering via Functional Entropy

AMPS uses instance-aware functional entropy to adaptively steer multimodal model modality preferences, improving control while minimizing inference errors.

Zihan Huang, Xintong Li, Rohan Surana, Tong Yu and 4 more

Atlanta Poster Session 2, Wed, Dec 9, 4:30 PM–7:30 PM, Hall C1 · Published 2026

Learning in Context, Guided by Choice: A Reward-Free Paradigm for Reinforcement Learning with Transformers

In-context preference-based RL trains transformers solely on preference feedback, achieving reward-free in-context generalization comparable to fully supervised methods.

Juncheng Dong, Moyang Guo, Bowen He, Ethan Fang and 2 more

Atlanta Poster Session 2, Wed, Dec 9, 4:30 PM–7:30 PM, Hall C1 · Published 2026

Leveraging Latent Visual Reasoning in Silence

Latent visual reasoning enhances multimodal training despite being largely unused at inference; attention-based reinforcement learning preserves its benefits by promoting latent-text interaction during training.

Dongyao Zhu, Zhen Wang, Xi Xiao, Han Jiang and 6 more

Atlanta Poster Session 2, Wed, Dec 9, 4:30 PM–7:30 PM, Hall C1 · Published 2026

Variational Trajectory Optimization of Anisotropic Diffusion Schedules

A variational framework learns matrix-valued anisotropic diffusion schedules and improves EDM across CIFAR-10, AFHQv2, FFHQ, and ImageNet-64.

Pengxi Liu, Zeyu M Li, Xiang Cheng

Atlanta Poster Session 1, Wed, Dec 9, 10:00 AM–1:00 PM, Hall C1 · Published 2026

DARTS: Targeting Prognostic Covariates in Budget-Constrained Sequential Experiments

DARTS sequentially acquires prognostic covariates within budget constraints to minimize variance while preserving valid causal inference coverage.

Kateryna Husar, Alexander Volfovsky

Sydney Poster Session 4, Wed, Dec 9, 5:00 PM–8:00 PM, Hall 1-4 · Published 2026

SimplexUQ: An Evaluation Framework and Benchmark for Conformal Uncertainty on Simplex-Valued Predictions

SimplexUQ benchmarks conformal wrappers on simplex-valued predictions, showing global calibration can hide severe under-coverage and no wrapper universally dominates across tasks and stratification maps.

Liang You, Hengyu Shi, Dongwen Ou

Atlanta Poster Session 2, Wed, Dec 9, 4:30 PM–7:30 PM, Hall C1 · Published 2026

Embedding Foundation Model Predictions in Discrete-Choice Models with Structural Guarantees

A two-stage adapter embeds foundation model predictions into a constrained multinomial logit, guaranteeing cost monotonicity and valid value-of-time estimates while improving choice accuracy by up to 12.8 percentage points.

Yingshuo Wang, Xian Sun, Yanhang Li, Zhichao Fan and 1 more

Atlanta Poster Session 5, Fri, Dec 11, 10:00 AM–1:00 PM, Hall C1 · Published 2026

LowRankArena: A Standardized Evaluation Platform for SVD-Based LLM Compression

LowRankArena standardizes SVD-based LLM compression evaluation and reveals that method rankings and speedups depend heavily on backbone and workload under aligned protocols.

Zishan Shao, Lixun Zhang, Kangning Cui, Wenhao Wu and 9 more

Atlanta Poster Session 1, Wed, Dec 9, 10:00 AM–1:00 PM, Hall C1 · Published 2026 · ▲ 1 on Hugging Face

MoBiQuant: Mixture-of-Bits Quantization for Token-Adaptive Any-Precision LLM

MoBiQuant uses recursive residual quantization and token-aware routing to adapt weight precision per token, improving any-precision LLM inference speed and memory.

Dongwei Wang, Jinhee Kim, Seokho Han, Denis Gudovskiy and 7 more

Atlanta Poster Session 2, Wed, Dec 9, 4:30 PM–7:30 PM, Hall C1 · Published 2026

LensVLM: Selective Context Expansion for Compressed Visual Representation of Text

LensVLM lets VLMs scan compressed rendered text and selectively expand only relevant regions via learned tools, maintaining near-full accuracy at 4.3x compression and outperforming baselines up to 10.1x across text QA benchmarks.

Roy Xie, Dan Friedman, Donghan Yu, Bowen Pan and 6 more

Sydney Poster Session 5, Thu, Dec 10, 10:00 AM–1:00 PM, Hall 1-4 · Published 2026 · ▲ 7 on Hugging Face

Learning Evidence Highlighting for Frozen LLMs

HiLight trains a lightweight actor via reinforcement learning to insert highlight tags around pivotal evidence spans in frozen LLM contexts, boosting reasoning without altering inputs or requiring evidence labels.

Shaoang Li, Yanhang Shi, Yufei Li, Mingfu Liang and 9 more

Atlanta Poster Session 2, Wed, Dec 9, 4:30 PM–7:30 PM, Hall C1 · Published 2026 · ▲ 5 on Hugging Face

Inertia-1: An Open Exploration of Wearable Motion Foundation Models

Inertia-1 explores wearable motion foundation models via 18.2M hours of accelerometer data, yielding state-of-the-art recipes and open design principles for diverse sensing tasks.

Zongzhe Xu, Aakarsh Anand, Sarah Jiang, Chuntung Zhuang and 3 more

Atlanta Poster Session 5, Fri, Dec 11, 10:00 AM–1:00 PM, Hall C1 · Published 2026 · ▲ 1 on Hugging Face · Code ★ 35

When Helpfulness Becomes Sycophancy: Sycophancy is a Boundary Failure Between Social Alignment and Epistemic Integrity in Large Language Models

Sycophancy is a boundary failure between social alignment and epistemic integrity, defined by three conditions involving cue expression, alignment shift, and compromised reasoning.

Jiechen Li, Catherine A Barry, Rishika Randev, Janet Chen and 2 more

Atlanta Poster Session 6, Fri, Dec 11, 4:30 PM–7:30 PM, Hall C1 · Published 2026

AstraFlow: Dataflow-Oriented Reinforcement Learning for Agentic LLMs

AstraFlow is a dataflow-oriented RL system for agentic LLMs that decouples rollout, dataflow, and training to enable multi-policy collaborative training with 2.7x faster training.

Haizhong Zheng, Yizhuo Di, Jiahui Wang, Shuowei Jin and 6 more

Atlanta Poster Session 6, Fri, Dec 11, 4:30 PM–7:30 PM, Hall C1 · Published 2026 · ▲ 15 on Hugging Face · Code ★ 105

Conservation Laws for Diffusion Models

Conservation laws express diffusion cross-entropy via local information-theoretic derivatives along noise paths, unifying discrete and continuous likelihoods and reducing training to marginal posterior learning.

Ziv Aharoni, Henry Pfister

Atlanta Poster Session 6, Fri, Dec 11, 4:30 PM–7:30 PM, Hall C1 · Published 2026

From Collapse to Improvement: Statistical Perspectives on the Evolutionary Dynamics of Iterative Training on Contaminated Sources

Statistical analysis shows iterative training on contaminated synthetic data avoids model collapse and recovers the true distribution with sufficient fresh samples and appropriate mixture weights.

Soham Bakshi, Sunrit Chakraborty

Atlanta Poster Session 3, Thu, Dec 10, 10:00 AM–1:00 PM, Hall C1 · Published 2026

FeatCal: Feature Calibration for Post-Merging Models

FeatCal reduces post-merging feature drift via layer-wise closed-form weight calibration without gradients, outperforming Surgery and ProbSurgery on CLIP and GLUE benchmarks.

Yanggan Gu, Shuo CAI, Zihao Wang, Wenjun Wang and 6 more

Sydney Poster Session 4, Wed, Dec 9, 5:00 PM–8:00 PM, Hall 1-4 · Published 2026

Simple KNN-Based Outlier Detection Achieves Robust Clustering

Simple KNN-based outlier removal achieves constant-factor robust k-means reductions with matching or better real-world clustering performance and speed.

Tianle Jiang, Yufa Zhou

Atlanta Poster Session 6, Fri, Dec 11, 4:30 PM–7:30 PM, Hall C1 · Published 2026

Population-Aligned Persona Generation for LLM-based Social Simulation

A framework generates population-aligned personas from social media via quality filtering, importance sampling, and task-specific adaptation, reducing bias in LLM social simulations.

Zhengyu Hu, Jianxun Lian, Zheyuan Xiao, Max Xiong and 9 more

Sydney Poster Session 3, Wed, Dec 9, 10:00 AM–1:00 PM, Hall 1-4 · Published 2026

Swimba: Switch Mamba Model Scales State Space Models

Swimba routes expert SSM streams via parameter-space MoE to scale selective state space model capacity without increasing recurrent state update costs. Under matched FLOPs, it achieves slightly better average performance with minor latency and throughput trade-offs.

Zhixu Du, Krishna Teja Chitty-Venkata, Murali Emani, Venkatram Vishwanath and 2 more

Atlanta Poster Session 3, Thu, Dec 10, 10:00 AM–1:00 PM, Hall C1 · Published 2026

Efficient Transferable Optimal Transport via Min-Sliced Transport Plans

Min-sliced transport plans transfer optimized slicers across related distributions with theoretical guarantees and efficient minibatch scaling for matching and generation.

Xinran Liu, Elaheh Akbari, Rocio Diaz Martin, Navid NaderiAlizadeh and 1 more

Atlanta Poster Session 5, Fri, Dec 11, 10:00 AM–1:00 PM, Hall C1 · Published 2026

A Multimodal Benchmark for Evaluating Cause-of-Death Inference Using Child Health and Mortality Data

This paper introduces a multimodal benchmark for cause-of-death inference in child mortality data, showing zero-shot language models synthesize unstructured medical evidence differently than supervised baselines.

Junhe Yang, Soumyakanti Pan, Hyun Seung Lim, YUE CHU and 17 more

Atlanta Poster Session 6, Fri, Dec 11, 4:30 PM–7:30 PM, Hall C1 · Published 2026

Corrective Diffusion Language Models

Standard diffusion language models lack reliable token correction, so a correction-oriented post-training principle improves iterative refinement and outperforms masked diffusion baselines.

Shuibai Zhang, Fred Peng, Yiheng Zhang, Jin Pan and 1 more

Atlanta Poster Session 3, Thu, Dec 10, 10:00 AM–1:00 PM, Hall C1 · Published 2026 · ▲ 1 on Hugging Face · Code ★ 17

Coupling Models for One-Step Discrete Generation

Coupling Models learn direct couplings between discrete sequences and Gaussian latents for one-step generation, reducing LM1B perplexity by 33%, Fly Brain FBD by 18%, and MNIST-Binary FID by 46%.

Fred Peng, Joey Bose, Anru Zhang, Alexander Tong

Atlanta Poster Session 1, Wed, Dec 9, 10:00 AM–1:00 PM, Hall C1 · Published 2026

Multistage Defer Trees for Hybrid Interpretability: If at First You Can't Succeed, Tree Again

Multistage Defer Trees use sequential sparse decision trees to route most samples through interpretable rules before deferring to black boxes, matching ensemble accuracy with high interpretability.

Zakk Heile, Hayden McTavish, Margo Seltzer, Cynthia Rudin

Sydney Poster Session 2, Tue, Dec 8, 5:00 PM–8:00 PM, Hall 1-4 · Published 2026

At FullTilt: Real-Time Open-Set 3D Macromolecule Detection Directly from Tilted 2D Projections

FullTilt detects open-set 3D macromolecules directly from 2D tilt-series via a tilt-series encoder, accelerating inference orders of magnitude while achieving state-of-the-art zero-shot results.

Ming-Yang Ho, Alberto Bartesaghi

Sydney Poster Session 4, Wed, Dec 9, 5:00 PM–8:00 PM, Hall 1-4 · Published 2026