Jianwen Xie
YOU?
Author Swipe
View article: Word Salad Chopper: Reasoning Models Waste A Ton Of Decoding Budget On Useless Repetitions, Self-Knowingly
Word Salad Chopper: Reasoning Models Waste A Ton Of Decoding Budget On Useless Repetitions, Self-Knowingly Open
Large Reasoning Models (LRMs) are often bottlenecked by the high cost of output tokens. We show that a significant portion of these tokens are useless self-repetitions - what we call "word salad" - that exhaust the decoding budget without …
View article: InstructPro: Natural Language Guided Ligand-Binding Protein Design
InstructPro: Natural Language Guided Ligand-Binding Protein Design Open
Designing ligand-binding proteins with precise functions is fundamental to advances in biology and chemistry, yet existing AI approaches are limited by scarce protein-ligand complex data. Meanwhile, abundant text descriptions of protein-li…
View article: Mitigating Non-IID Drift in Zeroth-Order Federated LLM Fine-Tuning with Transferable Sparsity
Mitigating Non-IID Drift in Zeroth-Order Federated LLM Fine-Tuning with Transferable Sparsity Open
Federated Learning enables collaborative fine-tuning of Large Language Models (LLMs) across decentralized Non-Independent and Identically Distributed (Non-IID) clients, but such models' massive parameter sizes lead to significant memory an…
View article: Latent Adaptive Planner for Dynamic Manipulation
Latent Adaptive Planner for Dynamic Manipulation Open
We present the Latent Adaptive Planner (LAP), a trajectory-level latent-variable policy for dynamic nonprehensile manipulation (e.g., box catching) that formulates planning as inference in a low-dimensional latent space and is learned effe…
View article: DEL-ToM: Inference-Time Scaling for Theory-of-Mind Reasoning via Dynamic Epistemic Logic
DEL-ToM: Inference-Time Scaling for Theory-of-Mind Reasoning via Dynamic Epistemic Logic Open
View article: Word Salad Chopper: Reasoning Models Waste A Ton Of Decoding Budget On Useless Repetitions, Self-Knowingly
Word Salad Chopper: Reasoning Models Waste A Ton Of Decoding Budget On Useless Repetitions, Self-Knowingly Open
View article: Latent Space Energy-based Neural ODEs
Latent Space Energy-based Neural ODEs Open
This paper introduces novel deep dynamical models designed to represent continuous-time sequences. Our approach employs a neural emission model to generate each data point in the time series through a non-linear transformation of a latent …
View article: Molecule Design by Latent Prompt Transformer
Molecule Design by Latent Prompt Transformer Open
This work explores the challenging problem of molecule design by framing it as a conditional generative modeling task, where target biological properties or desired chemical constraints serve as conditioning variables. We propose the Laten…
View article: Latent Plan Transformer for Trajectory Abstraction: Planning as Latent Space Inference
Latent Plan Transformer for Trajectory Abstraction: Planning as Latent Space Inference Open
In tasks aiming for long-term returns, planning becomes essential. We study generative modeling for planning with datasets repurposed from offline reinforcement learning. Specifically, we identify temporal consistency in the absence of ste…
View article: STANLEY: Stochastic Gradient Anisotropic Langevin Dynamics for Learning Energy-Based Models
STANLEY: Stochastic Gradient Anisotropic Langevin Dynamics for Learning Energy-Based Models Open
We propose in this paper, STANLEY, a STochastic gradient ANisotropic LangEvin dYnamics, for sampling high dimensional data. With the growing efficacy and potential of Energy-Based modeling, also known as non-normalized probabilistic modeli…
View article: Molecule Design by Latent Prompt Transformer
Molecule Design by Latent Prompt Transformer Open
This paper proposes a latent prompt Transformer model for solving challenging optimization problems such as molecule design, where the goal is to find molecules with optimal values of a target chemical or biological property that can be co…
View article: Progressive Energy-Based Cooperative Learning for Multi-Domain Image-to-Image Translation
Progressive Energy-Based Cooperative Learning for Multi-Domain Image-to-Image Translation Open
This paper studies a novel energy-based cooperative learning framework for multi-domain image-to-image translation. The framework consists of four components: descriptor, translator, style encoder, and style generator. The descriptor is a …
View article: Learning V1 Simple Cells with Vector Representation of Local Content and Matrix Representation of Local Motion
Learning V1 Simple Cells with Vector Representation of Local Content and Matrix Representation of Local Motion Open
This paper proposes a representational model for image pairs such as consecutive video frames that are related by local pixel displacements, in the hope that the model may shed light on motion perception in primary visual cortex (V1). The …
View article: Energy-Based Generative Cooperative Saliency Prediction
Energy-Based Generative Cooperative Saliency Prediction Open
Conventional saliency prediction models typically learn a deterministic mapping from an image to its saliency map, and thus fail to explain the subjective nature of human attention. In this paper, to model the uncertainty of visual salienc…
View article: Learning Generative Vision Transformer with Energy-Based Latent Space for Saliency Prediction
Learning Generative Vision Transformer with Energy-Based Latent Space for Saliency Prediction Open
Vision transformer networks have shown superiority in many computer vision tasks. In this paper, we take a step further by proposing a novel generative vision transformer with latent variables following an informative energy-based prior fo…
View article: Blind Image Super-resolution with Elaborate Degradation Modeling on Noise and Kernel
Blind Image Super-resolution with Elaborate Degradation Modeling on Noise and Kernel Open
While researches on model-based blind single image super-resolution (SISR) have achieved tremendous successes recently, most of them do not consider the image degradation sufficiently. Firstly, they always assume image noise obeys an indep…
View article: Unsupervised Single Image Super-resolution Under Complex Noise
Unsupervised Single Image Super-resolution Under Complex Noise Open
While the researches on single image super-resolution (SISR), especially equipped with deep neural networks (DNNs), have achieved tremendous successes recently, they still suffer from two major limitations. Firstly, the real image degradat…
View article: Learning Cycle-Consistent Cooperative Networks via Alternating MCMC Teaching for Unsupervised Cross-Domain Translation
Learning Cycle-Consistent Cooperative Networks via Alternating MCMC Teaching for Unsupervised Cross-Domain Translation Open
This paper studies the unsupervised cross-domain translation problem by proposing a generative framework, in which the probability distribution of each domain is represented by a generative cooperative network that consists of an energy-ba…
View article: Learning Energy-Based Model with Variational Auto-Encoder as Amortized Sampler
Learning Energy-Based Model with Variational Auto-Encoder as Amortized Sampler Open
Due to the intractable partition function, training energy-based models (EBMs) by maximum likelihood requires Markov chain Monte Carlo (MCMC) sampling to approximate the gradient of the Kullback-Leibler divergence between data and model di…
View article: Semi-Supervised Video Deraining with Dynamical Rain Generator
Semi-Supervised Video Deraining with Dynamical Rain Generator Open
While deep learning (DL)-based video deraining methods have achieved significant success recently, they still exist two major drawbacks. Firstly, most of them do not sufficiently model the characteristics of rain layers of rainy videos. In…
View article: Learning Cycle-Consistent Cooperative Networks via Alternating MCMC Teaching for Unsupervised Cross-Domain Translation
Learning Cycle-Consistent Cooperative Networks via Alternating MCMC Teaching for Unsupervised Cross-Domain Translation Open
This paper studies the unsupervised cross-domain translation problem by proposing a generative framework, in which the probability distribution of each domain is represented by a generative cooperative network that consists of an energy-ba…
View article: Learning Energy-Based Model with Variational Auto-Encoder as Amortized Sampler
Learning Energy-Based Model with Variational Auto-Encoder as Amortized Sampler Open
Due to the intractable partition function, training energy-based models (EBMs) by maximum likelihood requires Markov chain Monte Carlo (MCMC) sampling to approximate the gradient of the Kullback-Leibler divergence between data and model di…
View article: Learning Noise-Aware Encoder-Decoder from Noisy Labels by Alternating Back-Propagation for Saliency Detection
Learning Noise-Aware Encoder-Decoder from Noisy Labels by Alternating Back-Propagation for Saliency Detection Open
In this paper, we propose a noise-aware encoder-decoder framework to disentangle a clean saliency predictor from noisy training examples, where the noisy labels are generated by unsupervised handcrafted feature-based methods. The proposed …
View article: On Path Integration of Grid Cells: Group Representation and Isotropic Scaling
On Path Integration of Grid Cells: Group Representation and Isotropic Scaling Open
Understanding how grid cells perform path integration calculations remains a fundamental problem. In this paper, we conduct theoretical analysis of a general representation model of path integration by grid cells, where the 2D self-positio…
View article: On Path Integration of Grid Cells: Isotropic Metric, Conformal Embedding and Group Representation
On Path Integration of Grid Cells: Isotropic Metric, Conformal Embedding and Group Representation Open
The purpose of this paper is to understand how the grid cells may perform path integration calculations. We study a general representational model of path integration in which the self-position is represented by a vector formed by the acti…
View article: A Representational Model of Grid Cells Based on Matrix Lie Algebras.
A Representational Model of Grid Cells Based on Matrix Lie Algebras. Open
The grid cells in the mammalian medial entorhinal cortex exhibit striking hexagon firing patterns when the agent navigates in the open field. It is hypothesized that the grid cells are involved in path integral so that the agent is aware o…
View article: On Path Integration of Grid Cells: Group Representation and Isotropic\n Scaling
On Path Integration of Grid Cells: Group Representation and Isotropic\n Scaling Open
Understanding how grid cells perform path integration calculations remains a\nfundamental problem. In this paper, we conduct theoretical analysis of a\ngeneral representation model of path integration by grid cells, where the 2D\nself-posi…
View article: Motion-Based Generator Model: Unsupervised Disentanglement of Appearance, Trackable and Intrackable Motions in Dynamic Patterns
Motion-Based Generator Model: Unsupervised Disentanglement of Appearance, Trackable and Intrackable Motions in Dynamic Patterns Open
Dynamic patterns are characterized by complex spatial and motion patterns. Understanding dynamic patterns requires a disentangled representational model that separates the factorial components. A commonly used model for dynamic patterns is…
View article: Generative PointNet: Energy-Based Learning on Unordered Point Sets for 3D Generation, Reconstruction and Classification
Generative PointNet: Energy-Based Learning on Unordered Point Sets for 3D Generation, Reconstruction and Classification Open
We propose a generative model of unordered point sets, such as point clouds, in the form of an energy-based model, where the energy function is parameterized by an input-permutation-invariant bottom-up neural network. The energy function l…
View article: Representation Learning: A Statistical Perspective
Representation Learning: A Statistical Perspective Open
Learning representations of data is an important problem in statistics and machine learning. While the origin of learning representations can be traced back to factor analysis and multidimensional scaling in statistics, it has become a cen…