1 citations · 2 across the 19 of their papers we have counts for
13 papers · 1 filter
From Reasoning Traces to Reusable Modules: Understanding Compositional Generalization in Language Model Reasoning
Lingjing Kong, Xin Liu, Guangyi Chen +9
Post-training pipelines that combine supervised fine-tuning (SFT) with reinforcement learning (RL) have emerged as the key recipe for transforming large language models (LLMs) into…
MoRe: Modular Representations for Principled Continual Representation Learning on Sequential Data
Jiaqi Sun, Boyang Sun, Rasmy M. H. +2
Continual learning requires models to adapt to new data while preserving previously acquired knowledge. At its core, this challenge can be viewed as principled one-step adaptation:…
Beyond the Black Box: Identifiable Interpretation and Control in Generative Models via Causal Minimality
Lingjing Kong, Shaoan Xie, Guangyi Chen +4
Deep generative models, while revolutionizing fields like image and text generation, largely operate as opaque ``black boxes'', hindering human understanding, control, and alignmen…
Score-based Greedy Search for Structure Identification of Partially Observed Linear Causal Models
Xinshuai Dong, Ignavier Ng, Haoyue Dai +4
Identifying the structure of a partially observed causal system is essential to various scientific fields. Recent advances have focused on constraint-based causal discovery to solv…
LLM Interpretability with Identifiable Temporal-Instantaneous Representation
Xiangchen Song, Jiaqi Sun, Zijian Li +2
Despite Large Language Models' remarkable capabilities, understanding their internal representations remains challenging. Mechanistic interpretability tools such as sparse autoenco…
Position: Mechanistic Interpretability Should Prioritize Feature Consistency in SAEs
Xiangchen Song, Aashiq Muhamed, Yujia Zheng +5
Sparse Autoencoders (SAEs) are a prominent tool in mechanistic interpretability (MI) for decomposing neural network activations into interpretable features. However, the aspiration…
Causal Temporal Representation Learning with Nonstationary Sparse Transition
Xiangchen Song, Zijian Li, Guangyi Chen +4
Causal Temporal Representation Learning (Ctrl) methods aim to identify the temporal causal dynamics of complex nonstationary temporal sequences. Despite the success of existing Ctr…
On the Identification of Temporally Causal Representation with Instantaneous Dependence
Zijian Li, Yifan Shen, Kaitao Zheng +5
Temporally causal representation learning aims to identify the latent causal process from time series observations, but most methods require the assumption that the latent causal p…
Nonstationary Time Series Forecasting via Unknown Distribution Adaptation
Zijian Li, Ruichu Cai, Zhenhui Yang +6
As environments evolve, temporal distribution shifts can degrade time series forecasting performance. A straightforward solution is to adapt to nonstationary changes while preservi…
Calibration-then-Calculation: A Variance Reduced Metric Framework in Deep Click-Through Rate Prediction Models
Yewen Fan, Nian Si, Xiangchen Song +1
The adoption of deep learning across various fields has been extensive, yet there is a lack of focus on evaluating the performance of deep learning pipelines. Typically, with the i…
CaRiNG: Learning Temporal Causal Representation under Non-Invertible Generation Process
Guangyi Chen, Yifan Shen, Zhenhao Chen +5
Identifying the underlying time-delayed latent causal processes in sequential data is vital for grasping temporal dynamics and making downstream reasoning. While some recent method…
A Versatile Causal Discovery Framework to Allow Causally-Related Hidden Variables
Xinshuai Dong, Biwei Huang, Ignavier Ng +6
Most existing causal discovery methods rely on the assumption of no latent confounders, limiting their applicability in solving real-life problems. In this paper, we introduce a no…
Temporally Disentangled Representation Learning under Unknown Nonstationarity
Xiangchen Song, Weiran Yao, Yewen Fan +5
In unsupervised causal representation learning for sequential data with time-delayed latent causal influences, strong identifiability results for the disentanglement of causally-re…