An Empirical Study on Parameter-Efficient Fine-Tuning for MultiModal Large Language Models

Multimodal large language models (MLLMs) fine-tuned with multimodal instruction datasets have demonstrated remarkable capabilities in multimodal tasks. However, fine-tuning all parameters of MLLMs has become challenging as they usually contain billions of parameters. To address this issue, we study parameter-efficient fine-tuning (PEFT) methods for MLLMs. We aim to identify effective methods for enhancing the performance of MLLMs in scenarios where only a limited number of parameters are trained. This paper conducts empirical studies using four popular PEFT methods to fine-tune the LLM component of open-source MLLMs. We present a comprehensive analysis that encompasses various aspects, including the impact of PEFT methods on various models, parameters and location of the PEFT module, size of fine-tuning data, model stability based on PEFT methods, MLLM's generalization, and hallucination. We evaluated four PEFT methods on seven datasets from two different categories: unseen and seen datasets. Across all experiments, we show that the adapter is the best-performing PEFT method. At the same time, fine-tuning the connector layers leads to improved performance in most MLLMs. Code and data are available at https://github.com/alenai97/PEFT-MLLM.git.

UniArk: Improving Generalisation and Consistency for Factual Knowledge Extraction through Debiasing

Several recent papers have investigated the potential of language models as knowledge bases as well as the existence of severe biases when extracting factual knowledge. In this work, we focus on the factual probing performance over unseen prompts from tuning, and using a probabilistic view we show the inherent misalignment between pre-training and downstream tuning objectives in language models for probing knowledge. We hypothesize that simultaneously debiasing these objectives can be the key to generalisation over unseen prompts. We propose an adapter-based framework, UniArk, for generalised and consistent factual knowledge extraction through simple methods without introducing extra parameters. Extensive experiments show that UniArk can significantly improve the model's out-of-domain generalisation as well as consistency under various prompts. Additionally, we construct ParaTrex, a large-scale and diverse dataset for measuring the inconsistency and out-of-domain generation of models. Further, ParaTrex offers a reference method for constructing paraphrased datasets using large language models.

An Incremental Update Framework for Online Recommenders with Data-Driven Prior

Online recommenders have attained growing interest and created great revenue for businesses. Given numerous users and items, incremental update becomes a mainstream paradigm for learning large-scale models in industrial scenarios, where only newly arrived data within a sliding window is fed into the model, meeting the strict requirements of quick response. However, this strategy would be prone to overfitting to newly arrived data. When there exists a significant drift of data distribution, the long-term information would be discarded, which harms the recommendation performance. Conventional methods address this issue through native model-based continual learning methods, without analyzing the data characteristics for online recommenders. To address the aforementioned issue, we propose an incremental update framework for online recommenders with Data-Driven Prior (DDP), which is composed of Feature Prior (FP) and Model Prior (MP). The FP performs the click estimation for each specific value to enhance the stability of the training process. The MP incorporates previous model output into the current update while strictly following the Bayes rules, resulting in a theoretically provable prior for the robust update. In this way, both the FP and MP are well integrated into the unified framework, which is model-agnostic and can accommodate various advanced interaction models. Extensive experiments on two publicly available datasets as well as an industrial dataset demonstrate the superior performance of the proposed framework.

Data Contamination Issues in Brain-to-Text Decoding

Decoding non-invasive cognitive signals to natural language has long been the goal of building practical brain-computer interfaces (BCIs). Recent major milestones have successfully decoded cognitive signals like functional Magnetic Resonance Imaging (fMRI) and electroencephalogram (EEG) into text under open vocabulary setting. However, how to split the datasets for training, validating, and testing in cognitive signal decoding task still remains controversial. In this paper, we conduct systematic analysis on current dataset splitting methods and find the existence of data contamination largely exaggerates model performance. Specifically, first we find the leakage of test subjects' cognitive signals corrupts the training of a robust encoder. Second, we prove the leakage of text stimuli causes the auto-regressive decoder to memorize information in test set. The decoder generates highly accurate text not because it truly understands cognitive signals. To eliminate the influence of data contamination and fairly evaluate different models' generalization ability, we propose a new splitting method for different types of cognitive datasets (e.g. fMRI, EEG). We also test the performance of SOTA Brain-to-Text decoding models under the proposed dataset splitting paradigm as baselines for further research.

Parallel Ranking of Ads and Creatives in Real-Time Advertising Systems

"Creativity is the heart and soul of advertising services". Effective creatives can create a win-win scenario: advertisers can reach target users and achieve marketing objectives more effectively, users can more quickly find products of interest, and platforms can generate more advertising revenue. With the advent of AI-Generated Content, advertisers now can produce vast amounts of creative content at a minimal cost. The current challenge lies in how advertising systems can select the most pertinent creative in real-time for each user personally. Existing methods typically perform serial ranking of ads or creatives, limiting the creative module in terms of both effectiveness and efficiency. In this paper, we propose for the first time a novel architecture for online parallel estimation of ads and creatives ranking, as well as the corresponding offline joint optimization model. The online architecture enables sophisticated personalized creative modeling while reducing overall latency. The offline joint model for CTR estimation allows mutual awareness and collaborative optimization between ads and creatives. Additionally, we optimize the offline evaluation metrics for the implicit feedback sorting task involved in ad creative ranking. We conduct extensive experiments to compare ours with two state-of-the-art approaches. The results demonstrate the effectiveness of our approach in both offline evaluations and real-world advertising platforms online in terms of response time, CTR, and CPM.

Instances and Labels: Hierarchy-aware Joint Supervised Contrastive Learning for Hierarchical Multi-Label Text Classification

Hierarchical multi-label text classification (HMTC) aims at utilizing a label hierarchy in multi-label classification. Recent approaches to HMTC deal with the problem of imposing an over-constrained premise on the output space by using contrastive learning on generated samples in a semi-supervised manner to bring text and label embeddings closer. However, the generation of samples tends to introduce noise as it ignores the correlation between similar samples in the same batch. One solution to this issue is supervised contrastive learning, but it remains an underexplored topic in HMTC due to its complex structured labels. To overcome this challenge, we propose $\textbf{HJCL}$, a $\textbf{H}$ierarchy-aware $\textbf{J}$oint Supervised $\textbf{C}$ontrastive $\textbf{L}$earning method that bridges the gap between supervised contrastive learning and HMTC. Specifically, we employ both instance-wise and label-wise contrastive learning techniques and carefully construct batches to fulfill the contrastive learning objective. Extensive experiments on four multi-path HMTC datasets demonstrate that HJCL achieves promising results and the effectiveness of Contrastive Learning on HMTC.

OccupancyDETR: Making Semantic Scene Completion as Straightforward as Object Detection

Visual-based 3D semantic occupancy perception (also known as 3D semantic scene completion) is a new perception paradigm for robotic applications like autonomous driving. Compared with Bird's Eye View (BEV) perception, it extends the vertical dimension, significantly enhancing the ability of robots to understand their surroundings. However, due to this very reason, the computational demand for current 3D semantic occupancy perception methods generally surpasses that of BEV perception methods and 2D perception methods. We propose a novel 3D semantic occupancy perception method, OccupancyDETR, which consists of a DETR-like object detection module and a 3D occupancy decoder module. The integration of object detection simplifies our method structurally - instead of predicting the semantics of each voxels, it identifies objects in the scene and their respective 3D occupancy grids. This speeds up our method, reduces required resources, and leverages object detection algorithm, giving our approach notable performance on small objects. We demonstrate the effectiveness of our proposed method on the SemanticKITTI dataset, showcasing an mIoU of 23 and a processing speed of 6 frames per second, thereby presenting a promising solution for real-time 3D semantic scene completion.

BUCA: A Binary Classification Approach to Unsupervised Commonsense Question Answering

Unsupervised commonsense reasoning (UCR) is becoming increasingly popular as the construction of commonsense reasoning datasets is expensive, and they are inevitably limited in their scope. A popular approach to UCR is to fine-tune language models with external knowledge (e.g., knowledge graphs), but this usually requires a large number of training examples. In this paper, we propose to transform the downstream multiple choice question answering task into a simpler binary classification task by ranking all candidate answers according to their reasonableness. To this end, for training the model, we convert the knowledge graph triples into reasonable and unreasonable texts. Extensive experimental results show the effectiveness of our approach on various multiple choice question answering benchmarks. Furthermore, compared with existing UCR approaches using KGs, ours is less data hungry. Our code is available at https://github.com/probe2/BUCA.

JDRec: Practical Actor-Critic Framework for Online Combinatorial Recommender System

A combinatorial recommender (CR) system feeds a list of items to a user at a time in the result page, in which the user behavior is affected by both contextual information and items. The CR is formulated as a combinatorial optimization problem with the objective of maximizing the recommendation reward of the whole list. Despite its importance, it is still a challenge to build a practical CR system, due to the efficiency, dynamics, personalization requirement in online environment. In particular, we tear the problem into two sub-problems, list generation and list evaluation. Novel and practical model architectures are designed for these sub-problems aiming at jointly optimizing effectiveness and efficiency. In order to adapt to online case, a bootstrap algorithm forming an actor-critic reinforcement framework is given to explore better recommendation mode in long-term user interaction. Offline and online experiment results demonstrate the efficacy of proposed JDRec framework. JDRec has been applied in online JD recommendation, improving click through rate by 2.6% and synthetical value for the platform by 5.03%. We will publish the large-scale dataset used in this study to contribute to the research community.

IA-GCN: Interactive Graph Convolutional Network for Recommendation

Recently, Graph Convolutional Network (GCN) has become a novel state-of-art for Collaborative Filtering (CF) based Recommender Systems (RS). It is a common practice to learn informative user and item representations by performing embedding propagation on a user-item bipartite graph, and then provide the users with personalized item suggestions based on the representations. Despite effectiveness, existing algorithms neglect precious interactive features between user-item pairs in the embedding process. When predicting a user's preference for different items, they still aggregate the user tree in the same way, without emphasizing target-related information in the user neighborhood. Such a uniform aggregation scheme easily leads to suboptimal user and item representations, limiting the model expressiveness to some extent. In this work, we address this problem by building bilateral interactive guidance between each user-item pair and proposing a new model named IA-GCN (short for InterActive GCN). Specifically, when learning the user representation from its neighborhood, we assign higher attention weights to those neighbors similar to the target item. Correspondingly, when learning the item representation, we pay more attention to those neighbors resembling the target user. This leads to interactive and interpretable features, effectively distilling target-specific information through each graph convolutional operation. Our model is built on top of LightGCN, a state-of-the-art GCN model for CF, and can be combined with various GCN-based CF architectures in an end-to-end fashion. Extensive experiments on three benchmark datasets demonstrate the effectiveness and robustness of IA-GCN.