TEMPO: Temporal Preference Optimization of Video LLMs via Difficulty Scheduling and Pre-SFT Alignment

• URL: http://arxiv.org/abs/2503.16929v1
• Date: Fri, 21 Mar 2025 08:00:29 GMT
• Title: TEMPO: Temporal Preference Optimization of Video LLMs via Difficulty Scheduling and Pre-SFT Alignment
• Authors: Shicheng Li, Lei Li, Kun Ouyang, Shuhuai Ren, Yuanxin Liu, Yuanxing Zhang, Fuzheng Zhang, Lingpeng Kong, Qi Liu, Xu Sun,
• Abstract summary: TEMPO is a systematic framework that enhances temporal reasoning capabilities of Video Large Language Models.<n>Our approach consistently improves Video LLM performance across multiple benchmarks with a relatively small set of self-generated DPO data.<n>Our findings highlight our TEMPO as a scalable and efficient complement to SFT-based methods, paving the way for developing reliable Video LLMs.
• Score: 48.94844127553743
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Video Large Language Models (Video LLMs) have achieved significant success by leveraging a two-stage paradigm: pretraining on large-scale video-text data for vision-language alignment, followed by supervised fine-tuning (SFT) for task-specific capabilities. However, existing approaches struggle with temporal reasoning due to weak temporal correspondence in the data and reliance on the next-token prediction paradigm during training. To address these limitations, we propose TEMPO (TEMporal Preference Optimization), a systematic framework that enhances Video LLMs' temporal reasoning capabilities through Direct Preference Optimization (DPO). To facilitate this, we introduce an automated preference data generation pipeline that systematically constructs preference pairs by selecting videos that are rich in temporal information, designing video-specific perturbation strategies, and finally evaluating model responses on clean and perturbed video inputs. Our temporal alignment features two key innovations: curriculum learning which that progressively increases perturbation difficulty to improve model robustness and adaptability; and ``Pre-SFT Alignment'', applying preference optimization before instruction tuning to prioritize fine-grained temporal comprehension. Extensive experiments demonstrate that our approach consistently improves Video LLM performance across multiple benchmarks with a relatively small set of self-generated DPO data. We further analyze the transferability of DPO data across architectures and the role of difficulty scheduling in optimization. Our findings highlight our TEMPO as a scalable and efficient complement to SFT-based methods, paving the way for developing reliable Video LLMs.

Related papers

• Self-alignment of Large Video Language Models with Refined Regularized Preference Optimization [29.706347050700867]
  Large Video Language Models (LVLMs) struggle with fine-grained temporal understanding, hallucinate, and often make simple mistakes on even simple video question-answering tasks. We propose a self-alignment framework that enables LVLMs to learn from their own errors.
  arXiv  Detail & Related papers  (2025-04-16T13:43:56Z)
• VPO: Aligning Text-to-Video Generation Models with Prompt Optimization [80.86205966195593]
  Video generation models are typically trained on text-to-video pairs with highly detailed and carefully crafted descriptions. We introduce VPO, a principled framework that optimize prompts based on three core principles: harmlessness, accuracy, and helpfulness. Our experiments demonstrate that VPO significantly improves safety, alignment, and video quality compared to baseline methods.
  arXiv  Detail & Related papers  (2025-03-26T12:28:20Z)
• Temporal Preference Optimization for Long-Form Video Understanding [28.623353303256653]
  Temporal Preference Optimization (TPO) is a novel post-training framework designed to enhance the temporal grounding capabilities of video-LMMs.<n>TPO significantly enhances temporal understanding while reducing reliance on manually annotated data.<n>LLaVA-Video-TPO establishes itself as the leading 7B model on the Video-MME benchmark.
  arXiv  Detail & Related papers  (2025-01-23T18:58:03Z)
• Prompt-A-Video: Prompt Your Video Diffusion Model via Preference-Aligned LLM [54.2320450886902]
  Text-to-video models have made remarkable advancements through optimization on high-quality text-video pairs.<n>Current automatic methods for refining prompts encounter challenges such as Modality-Inconsistency, Cost-Discrepancy, and Model-Unaware.<n>We introduce Prompt-A-Video, which excels in crafting Video-Centric, Labor-Free and Preference-Aligned prompts tailored to specific video diffusion model.
  arXiv  Detail & Related papers  (2024-12-19T18:32:21Z)
• Bridging Vision and Language: Modeling Causality and Temporality in Video Narratives [0.0]
  We propose an enhanced framework that integrates a Causal-Temporal Reasoning Module into state-of-the-art LVLMs. CTRM comprises two key components: the Causal Dynamics (CDE) and the Temporal Learner (TRL) We design a multi-stage learning strategy to optimize the model, combining pre-training on large-scale video-text datasets.
  arXiv  Detail & Related papers  (2024-12-14T07:28:38Z)
• Free Video-LLM: Prompt-guided Visual Perception for Efficient Training-free Video LLMs [56.040198387038025]
  We present a novel prompt-guided visual perception framework (abbreviated as Free Video-LLM) for efficient inference of training-free video LLMs. Our method effectively reduces the number of visual tokens while maintaining high performance across multiple video question-answering benchmarks.
  arXiv  Detail & Related papers  (2024-10-14T12:35:12Z)
• Reference Trustable Decoding: A Training-Free Augmentation Paradigm for Large Language Models [79.41139393080736]
  Large language models (LLMs) have rapidly advanced and demonstrated impressive capabilities. In-Context Learning (ICL) and. Efficient Fine-Tuning (PEFT) are currently two mainstream methods for augmenting. LLMs to downstream tasks. We propose Reference Trustable Decoding (RTD), a paradigm that allows models to quickly adapt to new tasks without fine-tuning.
  arXiv  Detail & Related papers  (2024-09-30T10:48:20Z)
• Self-Augmented Preference Optimization: Off-Policy Paradigms for Language Model Alignment [104.18002641195442]
  We introduce Self-Augmented Preference Optimization (SAPO), an effective and scalable training paradigm that does not require existing paired data. Building on the self-play concept, which autonomously generates negative responses, we further incorporate an off-policy learning pipeline to enhance data exploration and exploitation.
  arXiv  Detail & Related papers  (2024-05-31T14:21:04Z)
• ST-LLM: Large Language Models Are Effective Temporal Learners [58.79456373423189]
  Large Language Models (LLMs) have showcased impressive capabilities in text comprehension and generation. How to effectively encode and understand videos in video-based dialogue systems remains to be solved. We propose ST-LLM, an effective video-LLM baseline with spatial-temporal sequence modeling inside LLM.
  arXiv  Detail & Related papers  (2024-03-30T10:11:26Z)
• Unleashing the Potential of Large Language Models as Prompt Optimizers: Analogical Analysis with Gradient-based Model Optimizers [108.72225067368592]
  We propose a novel perspective to investigate the design of large language models (LLMs)-based prompts.<n>We identify two pivotal factors in model parameter learning: update direction and update method.<n>We develop a capable Gradient-inspired Prompt-based GPO.
  arXiv  Detail & Related papers  (2024-02-27T15:05:32Z)
• LLM4TS: Aligning Pre-Trained LLMs as Data-Efficient Time-Series Forecasters [11.796765525301051]
  We propose a framework for time-series forecasting with pre-trained Large Language Models (LLMs)<n>LLM4TS consists of a two-stage fine-tuning strategy to align LLMs with the nuances of time-series data, and the forecasting fine-tuning stage for downstream time-series forecasting tasks.<n>Our framework features a novel two-level aggregation method that integrates multi-scale temporal data within pre-trained LLMs, enhancing their ability to interpret time-specific information.
  arXiv  Detail & Related papers  (2023-08-16T16:19:50Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.