11 citations · 17 across the 26 of their papers we have counts for
27 papers
Mid-Training Language Models on Raw Video
Jaedong Hwang, Xiaoqian Shen, Ernie Chang +9
Multimodal large language models learn mostly from paired image-text data or annotated video, and raw web video is rarely used to further train an existing language model. We study…
The Safety Relay in Roleplay Jailbreaks: A Component-Resolved Causal Analysis of Harm Recognition and Refusal
Md Mokarram Chowdhury, Ernie Chang, Yang Li
Large language models are trained to follow instructions while refusing harmful requests. Jailbreaks exploit this balance to elicit content a model would ordinarily reject. Rolepla…
MobileMoE: Scaling On-Device Mixture of Experts
Yanbei Chen, Hanxian Huang, Ernie Chang +5
Mixture-of-Experts (MoE) has become the de facto architecture for hundred-billion-parameter language models, yet its advantages at sub-billion scales for on-device deployment remai…
Importance-Guided Basis Selection for Low-Rank Decomposition of Large Language Models
Daniel Agyei Asante, Ernie Chang, Yang Li
Low-rank decomposition is a compelling approach for compressing large language models, but its effectiveness hinges on selecting which singular-vector bases to retain for a target…
Exploring Audio Hallucination in Egocentric Video Understanding
Ashish Seth, Xinhao Mei, Changsheng Zhao +9
Egocentric videos provide a distinctive setting in which sound serves as crucial cues to understand user activities and surroundings, particularly when visual information is unstab…
Neural Computers
Mingchen Zhuge, Changsheng Zhao, Haozhe Liu +16
We propose a new frontier: Neural Computers (NCs) that unify computation, memory, and I/O of traditional computers in a learned runtime state. Our long-term goal is the Completely…
RPRA: Predicting an LLM-Judge for Efficient but Performant Inference
Dylan R. Ashley, Gaël Le Lan, Changsheng Zhao +7
Large language models (LLMs) face a fundamental trade-off between computational efficiency (e.g., number of parameters) and output quality, especially when deployed on computationa…
dTRPO: Trajectory Reduction in Policy Optimization of Diffusion Large Language Models
Wenxuan Zhang, Lemeng Wu, Changsheng Zhao +11
Diffusion Large Language Models (dLLMs) introduce a new paradigm for language generation, which in turn presents new challenges for aligning them with human preferences. In this wo…
Short Data, Long Context: Distilling Positional Knowledge in Transformers
Patrick Huber, Ernie Chang, Chinnadhurai Sankar +4
Extending the context window of language models typically requires expensive long-context pre-training, posing significant challenges for both training efficiency and data collecti…
EgoAVU: Egocentric Audio-Visual Understanding
Ashish Seth, Xinhao Mei, Changsheng Zhao +9
Understanding egocentric videos plays a vital role for embodied intelligence. Recent multi-modal large language models (MLLMs) can accept both visual and audio inputs. However, due…
VideoAuto-R1: Video Auto Reasoning via Thinking Once, Answering Twice
Shuming Liu, Mingchen Zhuge, Changsheng Zhao +20
Chain-of-thought (CoT) reasoning has emerged as a powerful tool for multimodal large language models on video understanding tasks. However, its necessity and advantages over direct…
MobileLLM-Pro Technical Report
Patrick Huber, Ernie Chang, Wei Wen +16
Efficient on-device language models around 1 billion parameters are essential for powering low-latency AI applications on mobile and wearable devices. However, achieving strong per…
MobileLLM-R1: Exploring the Limits of Sub-Billion Language Model Reasoners with Open Training Recipes
Changsheng Zhao, Ernie Chang, Zechun Liu +8
The paradigm shift in large language models (LLMs) from instinctive responses to chain-of-thought (CoT) reasoning has fueled two prevailing assumptions: (1) reasoning capabilities…
IMPACT: Importance-Aware Activation Space Reconstruction
Md Mokarram Chowdhury, Daniel Agyei Asante, Ernie Chang +1
Large language models (LLMs) achieve strong performance across diverse domains but remain difficult to deploy in resource-constrained environments due to their size. Low-rank compr…
AutoMixer: Checkpoint Artifacts as Automatic Data Mixers
Ernie Chang, Yang Li, Patrick Huber +4
In language model training, it is desirable to equip models with capabilities from various tasks. However, it is not clear how to directly obtain the right data mixtures for these…
Self-Vocabularizing Training for Neural Machine Translation
Pin-Jie Lin, Ernie Chang, Yangyang Shi +1
Past vocabulary learning techniques identify relevant vocabulary before training, relying on statistical and entropy-based assumptions that largely neglect the role of model traini…
CoSMoEs: Compact Sparse Mixture of Experts
Patrick Huber, Akshat Shrivastava, Ernie Chang +3
Sparse Mixture of Expert (MoE) models are popular foundational architectures at large scale, however, under-explored at smaller sizes. Here, we show how to enable Compact Sparse Mi…
Agent-as-a-Judge: Evaluate Agents with Agents
Mingchen Zhuge, Changsheng Zhao, Dylan Ashley +10
Contemporary evaluation techniques are inadequate for agentic systems. These approaches either focus exclusively on final outcomes -- ignoring the step-by-step nature of agentic sy…
Scaling Parameter-Constrained Language Models with Quality Data
Ernie Chang, Matteo Paltenghi, Yang Li +7
Scaling laws in language modeling traditionally quantify training loss as a function of dataset size and model parameters, providing compute-optimal estimates but often neglecting…
Target-Aware Language Modeling via Granular Data Sampling
Ernie Chang, Pin-Jie Lin, Yang Li +6
Language model pretraining generally targets a broad range of use cases and incorporates data from diverse sources. However, there are instances where we desire a model that excels…
High Fidelity Text-Guided Music Editing via Single-Stage Flow Matching
Gael Le Lan, Bowen Shi, Zhaoheng Ni +9
We introduce MelodyFlow, an efficient text-controllable high-fidelity music generation and editing model. It operates on continuous latent representations from a low frame rate 48…
Navigating the Minefield of MT Beam Search in Cascaded Streaming Speech Translation
Rastislav Rabatin, Frank Seide, Ernie Chang
We adapt the well-known beam-search algorithm for machine translation to operate in a cascaded real-time speech translation system. This proved to be more complex than initially an…
Basis Selection: Low-Rank Decomposition of Pretrained Large Language Models for Target Applications
Yang Li, Daniel Agyei Asante, Changsheng Zhao +3
Large language models (LLMs) significantly enhance the performance of various applications, but they are computationally intensive and energy-demanding. This makes it challenging t…
MobileLLM: Optimizing Sub-billion Parameter Language Models for On-Device Use Cases
Zechun Liu, Changsheng Zhao, Forrest Iandola +9
This paper addresses the growing need for efficient large language models (LLMs) on mobile devices, driven by increasing cloud costs and latency concerns. We focus on designing top…