Yuandong Tian

38

Papers

406

Total Citations

Papers (38)

ELF: An Extensive, Lightweight and Flexible Research Platform for Real-time Strategy Games

NeurIPS 2017arXiv

AdvPrompter: Fast Adaptive Adversarial Prompting for LLMs

NaturalReasoning: Reasoning in the Wild with 2.8M Challenging Questions

JoMA: Demystifying Multilayer Transformers via Joint Dynamics of MLP and Attention

Token Assorted: Mixing Latent and Text Tokens for Improved Language Model Reasoning

From Low Rank Gradient Subspace Stabilization to Low-Rank Weights: Observations, Theories, and Applications

LoCoCo: Dropping In Convolutions for Long Context Compression

MobileLLM: Optimizing Sub-billion Parameter Language Models for On-Device Use Cases

GenCO: Generating Diverse Designs with Combinatorial Constraints

TravelPlanner: A Benchmark for Real-World Planning with Language Agents

GaLore: Memory-Efficient LLM Training by Gradient Low-Rank Projection

Contrastive Predict-and-Search for Mixed Integer Linear Programs

Semantic Amodal Segmentation

FBNet: Hardware-Aware Efficient ConvNet Design via Differentiable Neural Architecture Search

FBNetV2: Differentiable Neural Architecture Search for Spatial and Channel Dimensions

FP-NAS: Fast Probabilistic Neural Architecture Search

FBNetV3: Joint Architecture-Recipe Search Using Predictor Pretraining

On the Importance of Asymmetry for Siamese Representation Learning

Bayesian Relational Memory for Semantic Visual Navigation

Param$\Delta$ for Direct Mixing: Post-Train Large Language Model At Zero Cost

Coda: An End-to-End Neural Program Decompiler

Learning to Perform Local Rewriting for Combinatorial Optimization

Hierarchical Decision Making by Generating and Following Natural Language Instructions

One ticket to win them all: generalizing lottery ticket initializations across datasets and optimizers

Learning Search Space Partition for Black-box Optimization using Monte Carlo Tree Search

Joint Policy Search for Multi-agent Collaboration with Imperfect Information

Learning Space Partitions for Path Planning

MADE: Exploration via Maximizing Deviation from Explored Regions

Latent Execution for Neural Program Synthesis Beyond Domain-Specific Languages

NovelD: A Simple yet Effective Exploration Criterion

DreamShard: Generalizable Embedding Table Placement for Recommender Systems

Understanding Deep Contrastive Learning via Coordinate-wise Optimization

Landscape Surrogate: Learning Decision Losses for Mathematical Optimization Under Partial Information

H2O: Heavy-Hitter Oracle for Efficient Generative Inference of Large Language Models

Scan and Snap: Understanding Training Dynamics and Token Composition in 1-layer Transformer

An Analytical Formula of Population Gradient for two-layered ReLU network and its Applications in Convergence and Critical Point Analysis

Gradient Descent Learns One-hidden-layer CNN: Don’t be Afraid of Spurious Local Minima

ELF OpenGo: an analysis and open reimplementation of AlphaZero