The recent surge in Large Language Models (LLMs) has garnered significant attention across numerous fields. Fine-tuning is often required to fit general LLMs for a specific domain, like the web-based healthcare system. However, two problems arise during fine-tuning LLMs for medical applications. One is the task variety problem, which involves distinct tasks in real-world medical scenarios. The variety often leads to sub-optimal fine-tuning for data imbalance and seesaw problems. Besides, the large amount of parameters in LLMs leads to huge time and computation consumption by fine-tuning. To address these two problems, we propose a novel parameter efficient fine-tuning framework for multi-task medical applications, dubbed as MOELoRA. The designed framework aims to absorb both the benefits of mixture-of-expert (MOE) for multi-task learning and low-rank adaptation (LoRA) for parameter efficient fine-tuning. For unifying MOE and LoRA, we devise multiple experts as the trainable parameters, where each expert consists of a pair of low-rank matrices to retain the small size of trainable parameters. Then, a task-motivated gate function for all MOELoRA layers is proposed, which can control the contributions of each expert and produce distinct parameters for various tasks. We conduct experiments on a multi-task medical dataset, indicating MOELoRA outperforms the existing parameter efficient fine-tuning methods. The code is available online.
Gait analysis is proven to be a reliable way to perform person identification without relying on subject cooperation. Walking is a biometric that does not significantly change in short periods of time and can be regarded as unique to each person. So far, the study of gait analysis focused mostly on identification and demographics estimation, without considering many of the pedestrian attributes that appearance-based methods rely on. In this work, alongside gait-based person identification, we explore pedestrian attribute identification solely from movement patterns. We propose DenseGait, the largest dataset for pretraining gait analysis systems containing 217K anonymized tracklets, annotated automatically with 42 appearance attributes. DenseGait is constructed by automatically processing video streams and offers the full array of gait covariates present in the real world. We make the dataset available to the research community. Additionally, we propose GaitFormer, a transformer-based model that after pretraining in a multi-task fashion on DenseGait, achieves 92.5% accuracy on CASIA-B and 85.33% on FVG, without utilizing any manually annotated data. This corresponds to a +14.2% and +9.67% accuracy increase compared to similar methods. Moreover, GaitFormer is able to accurately identify gender information and a multitude of appearance attributes utilizing only movement patterns. The code to reproduce the experiments is made publicly.
A well-studied generalization of the standard online convex optimization (OCO) framework is constrained online convex optimization (COCO). In COCO, on every round, a convex cost function and a convex constraint function are revealed to the learner after it chooses the action for that round. The objective is to design an online learning policy that simultaneously achieves a small regret while ensuring a small cumulative constraint violation (CCV) against an adaptive adversary interacting over a horizon of length $T$. A long-standing open question in COCO is whether an online policy can simultaneously achieve $O(\sqrt{T})$ regret and $\tilde{O}(\sqrt{T})$ CCV without any restrictive assumptions. For the first time, we answer this in the affirmative and show that a simple first-order policy can simultaneously achieve these bounds. Furthermore, in the case of strongly convex cost and convex constraint functions, the regret guarantee can be improved to $O(\log T)$ while keeping the CCV bound the same as above. We establish these results by effectively combining adaptive OCO policies as a blackbox with Lyapunov optimization - a classic tool from control theory. Surprisingly, the analysis is short and elegant.
In Conversational Recommendation System (CRS), an agent is asked to recommend a set of items to users within natural language conversations. To address the need for both conversational capability and personalized recommendations, prior works have utilized separate recommendation and dialogue modules. However, such approach inevitably results in a discrepancy between recommendation results and generated responses. To bridge the gap, we propose a multi-task learning for a unified CRS, where a single model jointly learns both tasks via Contextualized Knowledge Distillation (ConKD). We introduce two versions of ConKD: hard gate and soft gate. The former selectively gates between two task-specific teachers, while the latter integrates knowledge from both teachers. Our gates are computed on-the-fly in a context-specific manner, facilitating flexible integration of relevant knowledge. Extensive experiments demonstrate that our single model significantly improves recommendation performance while enhancing fluency, and achieves comparable results in terms of diversity.
Simultaneous sequence generation is a pivotal task for real-time scenarios, such as streaming speech recognition, simultaneous machine translation and simultaneous speech translation, where the target sequence is generated while receiving the source sequence. The crux of achieving high-quality generation with low latency lies in identifying the optimal moments for generating, accomplished by learning a mapping between the source and target sequences. However, existing methods often rely on task-specific heuristics for different sequence types, limiting the model's capacity to adaptively learn the source-target mapping and hindering the exploration of multi-task learning for various simultaneous tasks. In this paper, we propose a unified segment-to-segment framework (Seg2Seg) for simultaneous sequence generation, which learns the mapping in an adaptive and unified manner. During the process of simultaneous generation, the model alternates between waiting for a source segment and generating a target segment, making the segment serve as the natural bridge between the source and target. To accomplish this, Seg2Seg introduces a latent segment as the pivot between source to target and explores all potential source-target mappings via the proposed expectation training, thereby learning the optimal moments for generating. Experiments on multiple simultaneous generation tasks demonstrate that Seg2Seg achieves state-of-the-art performance and exhibits better generality across various tasks.
Representational spaces learned via language modeling are fundamental to Natural Language Processing (NLP), however there has been limited understanding regarding how and when during training various types of linguistic information emerge and interact. Leveraging a novel information theoretic probing suite, which enables direct comparisons of not just task performance, but their representational subspaces, we analyze nine tasks covering syntax, semantics and reasoning, across 2M pre-training steps and five seeds. We identify critical learning phases across tasks and time, during which subspaces emerge, share information, and later disentangle to specialize. Across these phases, syntactic knowledge is acquired rapidly after 0.5% of full training. Continued performance improvements primarily stem from the acquisition of open-domain knowledge, while semantics and reasoning tasks benefit from later boosts to long-range contextualization and higher specialization. Measuring cross-task similarity further reveals that linguistically related tasks share information throughout training, and do so more during the critical phase of learning than before or after. Our findings have implications for model interpretability, multi-task learning, and learning from limited data.
When a number of similar tasks have to be learned simultaneously, multi-task learning (MTL) models can attain significantly higher accuracy than single-task learning (STL) models. However, the advantage of MTL depends on various factors, such as the similarity of the tasks, the sizes of the datasets, and so on; in fact, some tasks might not benefit from MTL and may even incur a loss of accuracy compared to STL. Hence, the question arises: which tasks should be learned together? Domain experts can attempt to group tasks together following intuition, experience, and best practices, but manual grouping can be labor-intensive and far from optimal. In this paper, we propose a novel automated approach for task grouping. First, we study the affinity of tasks for MTL using four benchmark datasets that have been used extensively in the MTL literature, focusing on neural network-based MTL models. We identify inherent task features and STL characteristics that can help us to predict whether a group of tasks should be learned together using MTL or if they should be learned independently using STL. Building on this predictor, we introduce a randomized search algorithm, which employs the predictor to minimize the number of MTL trainings performed during the search for task groups. We demonstrate on the four benchmark datasets that our predictor-driven search approach can find better task groupings than existing baseline approaches.
The construction of large open knowledge bases (OKBs) is integral to many applications in the field of mobile computing. Noun phrases and relational phrases in OKBs often suffer from redundancy and ambiguity, which calls for the investigation on OKB canonicalization. However, in order to meet the requirements of some privacy protection regulations and to ensure the timeliness of the data, the canonicalized OKB often needs to remove some sensitive information or outdated data. The machine unlearning in OKB canonicalization is an excellent solution to the above problem. Current solutions address OKB canonicalization by devising advanced clustering algorithms and using knowledge graph embedding (KGE) to further facilitate the canonicalization process. Effective schemes are urgently needed to fully synergise machine unlearning with clustering and KGE learning. To this end, we put forward a multi-task unlearning framework, namely MulCanon, to tackle machine unlearning problem in OKB canonicalization. Specifically, the noise characteristics in the diffusion model are utilized to achieve the effect of machine unlearning for data in OKB. MulCanon unifies the learning objectives of diffusion model, KGE and clustering algorithms, and adopts a two-step multi-task learning paradigm for training. A thorough experimental study on popular OKB canonicalization datasets validates that MulCanon achieves advanced machine unlearning effects.
The landscape of publicly available vision foundation models (VFMs), such as CLIP and Segment Anything Model (SAM), is expanding rapidly. VFMs are endowed with distinct capabilities stemming from their pre-training objectives. For instance, CLIP excels in semantic understanding, while SAM specializes in spatial understanding for segmentation. In this work, we introduce a simple recipe to efficiently merge VFMs into a unified model that absorbs their expertise. Our method integrates techniques of multi-task learning, continual learning, and distillation. Further, it demands significantly less computational cost compared to traditional multi-task training from scratch, and it only needs a small fraction of the pre-training datasets that were initially used to train individual models. By applying our method to SAM and CLIP, we obtain SAM-CLIP: a unified model that combines the capabilities of SAM and CLIP into a single vision transformer. Compared with deploying SAM and CLIP independently, our merged model, SAM-CLIP, reduces storage and compute costs for inference, making it well-suited for edge device applications. We show that SAM-CLIP not only retains the foundational strengths of SAM and CLIP, but also introduces synergistic functionalities, notably in zero-shot semantic segmentation, where SAM-CLIP establishes new state-of-the-art results on 5 benchmarks. It outperforms previous models that are specifically designed for this task by a large margin, including +6.8% and +5.9% mean IoU improvement on Pascal-VOC and COCO-Stuff datasets, respectively.
The generalisation capacity of Multi-Task Learning (MTL) suffers when unrelated tasks negatively impact each other by updating shared parameters with conflicting gradients. This is known as negative transfer and leads to a drop in MTL accuracy compared to single-task learning (STL). Lately, there has been a growing focus on the fairness of MTL models, requiring the optimization of both accuracy and fairness for individual tasks. Analogously to negative transfer for accuracy, task-specific fairness considerations might adversely affect the fairness of other tasks when there is a conflict of fairness loss gradients between the jointly learned tasks - we refer to this as Bias Transfer. To address both negative- and bias-transfer in MTL, we propose a novel method called FairBranch, which branches the MTL model by assessing the similarity of learned parameters, thereby grouping related tasks to alleviate negative transfer. Moreover, it incorporates fairness loss gradient conflict correction between adjoining task-group branches to address bias transfer within these task groups. Our experiments on tabular and visual MTL problems show that FairBranch outperforms state-of-the-art MTLs on both fairness and accuracy.
Recent research has demonstrated the efficacy of pre-training graph neural networks (GNNs) to capture the transferable graph semantics and enhance the performance of various downstream tasks. However, the semantic knowledge learned from pretext tasks might be unrelated to the downstream task, leading to a semantic gap that limits the application of graph pre-training. To reduce this gap, traditional approaches propose hybrid pre-training to combine various pretext tasks together in a multi-task learning fashion and learn multi-grained knowledge, which, however, cannot distinguish tasks and results in some transferable task-specific knowledge distortion by each other. Moreover, most GNNs cannot distinguish nodes located in different parts of the graph, making them fail to learn position-specific knowledge and lead to suboptimal performance. In this work, inspired by the prompt-based tuning in natural language processing, we propose a unified framework for graph hybrid pre-training which injects the task identification and position identification into GNNs through a prompt mechanism, namely multi-task graph dual prompt (ULTRA-DP). Based on this framework, we propose a prompt-based transferability test to find the most relevant pretext task in order to reduce the semantic gap. To implement the hybrid pre-training tasks, beyond the classical edge prediction task (node-node level), we further propose a novel pre-training paradigm based on a group of $k$-nearest neighbors (node-group level). The combination of them across different scales is able to comprehensively express more structural semantics and derive richer multi-grained knowledge. Extensive experiments show that our proposed ULTRA-DP can significantly enhance the performance of hybrid pre-training methods and show the generalizability to other pre-training tasks and backbone architectures.
In this paper, we propose CLMSM, a domain-specific, continual pre-training framework, that learns from a large set of procedural recipes. CLMSM uses a Multi-Task Learning Framework to optimize two objectives - a) Contrastive Learning using hard triplets to learn fine-grained differences across entities in the procedures, and b) a novel Mask-Step Modelling objective to learn step-wise context of a procedure. We test the performance of CLMSM on the downstream tasks of tracking entities and aligning actions between two procedures on three datasets, one of which is an open-domain dataset not conforming with the pre-training dataset. We show that CLMSM not only outperforms baselines on recipes (in-domain) but is also able to generalize to open-domain procedural NLP tasks.
Cross-Domain Recommendation (CDR) stands as a pivotal technology addressing issues of data sparsity and cold start by transferring general knowledge from the source to the target domain. However, existing CDR models suffer limitations in adaptability across various scenarios due to their inherent complexity. To tackle this challenge, recent advancements introduce universal CDR models that leverage shared embeddings to capture general knowledge across domains and transfer it through "Multi-task Learning" or "Pre-train, Fine-tune" paradigms. However, these models often overlook the broader structural topology that spans domains and fail to align training objectives, potentially leading to negative transfer. To address these issues, we propose a motif-based prompt learning framework, MOP, which introduces motif-based shared embeddings to encapsulate generalized domain knowledge, catering to both intra-domain and inter-domain CDR tasks. Specifically, we devise three typical motifs: butterfly, triangle, and random walk, and encode them through a Motif-based Encoder to obtain motif-based shared embeddings. Moreover, we train MOP under the "Pre-training \& Prompt Tuning" paradigm. By unifying pre-training and recommendation tasks as a common motif-based similarity learning task and integrating adaptable prompt parameters to guide the model in downstream recommendation tasks, MOP excels in transferring domain knowledge effectively. Experimental results on four distinct CDR tasks demonstrate the effectiveness of MOP than the state-of-the-art models.
In multi-task reinforcement learning (RL) under Markov decision processes (MDPs), the presence of shared latent structures among multiple MDPs has been shown to yield significant benefits to the sample efficiency compared to single-task RL. In this paper, we investigate whether such a benefit can extend to more general sequential decision making problems, such as partially observable MDPs (POMDPs) and more general predictive state representations (PSRs). The main challenge here is that the large and complex model space makes it hard to identify what types of common latent structure of multi-task PSRs can reduce the model complexity and improve sample efficiency. To this end, we posit a joint model class for tasks and use the notion of $\eta$-bracketing number to quantify its complexity; this number also serves as a general metric to capture the similarity of tasks and thus determines the benefit of multi-task over single-task RL. We first study upstream multi-task learning over PSRs, in which all tasks share the same observation and action spaces. We propose a provably efficient algorithm UMT-PSR for finding near-optimal policies for all PSRs, and demonstrate that the advantage of multi-task learning manifests if the joint model class of PSRs has a smaller $\eta$-bracketing number compared to that of individual single-task learning. We also provide several example multi-task PSRs with small $\eta$-bracketing numbers, which reap the benefits of multi-task learning. We further investigate downstream learning, in which the agent needs to learn a new target task that shares some commonalities with the upstream tasks via a similarity constraint. By exploiting the learned PSRs from the upstream, we develop a sample-efficient algorithm that provably finds a near-optimal policy.
In end-to-end autonomous driving, the utilization of existing sensor fusion techniques and navigational control methods for imitation learning proves inadequate in challenging situations that involve numerous dynamic agents. To address this issue, we introduce LeTFuser, a lightweight transformer-based algorithm for fusing multiple RGB-D camera representations. To perform perception and control tasks simultaneously, we utilize multi-task learning. Our model comprises of two modules, the first being the perception module that is responsible for encoding the observation data obtained from the RGB-D cameras. Our approach employs the Convolutional vision Transformer (CvT) \cite{wu2021cvt} to better extract and fuse features from multiple RGB cameras due to local and global feature extraction capability of convolution and transformer modules, respectively. Encoded features combined with static and dynamic environments are later employed by our control module to predict waypoints and vehicular controls (e.g. steering, throttle, and brake). We use two methods to generate the vehicular controls levels. The first method uses a PID algorithm to follow the waypoints on the fly, whereas the second one directly predicts the control policy using the measurement features and environmental state. We evaluate the model and conduct a comparative analysis with recent models on the CARLA simulator using various scenarios, ranging from normal to adversarial conditions, to simulate real-world scenarios. Our method demonstrated better or comparable results with respect to our baselines in term of driving abilities. The code is available at \url{https://github.com/pagand/e2etransfuser/tree/cvpr-w} to facilitate future studies.
Machine learning strategies like multi-task learning, meta-learning, and transfer learning enable efficient adaptation of machine learning models to specific applications in healthcare, such as prediction of various diseases, by leveraging generalizable knowledge across large datasets and multiple domains. In particular, Bayesian meta-learning methods pool data across related prediction tasks to learn prior distributions for model parameters, which are then used to derive models for specific tasks. However, inter- and intra-task variability due to disease heterogeneity and other patient-level differences pose challenges of negative transfer during shared learning and poor generalizability to new patients. We introduce a novel Bayesian meta-learning approach that aims to address this in two key settings: (1) predictions for new patients (same population as the training set) and (2) adapting to new patient populations. Our main contribution is in modeling similarity between causal mechanisms of the tasks, for (1) mitigating negative transfer during training and (2) fine-tuning that pools information from tasks that are expected to aid generalizability. We propose an algorithm for implementing this approach for Bayesian deep learning, and apply it to a case study for stroke prediction tasks using electronic health record data. Experiments for the UK Biobank dataset as the training population demonstrated significant generalizability improvements compared to standard meta-learning, non-causal task similarity measures, and local baselines (separate models for each task). This was assessed for a variety of tasks that considered both new patients from the training population (UK Biobank) and a new population (FinnGen).
In recent years, DL has developed rapidly, and personalized services are exploring using DL algorithms to improve the performance of the recommendation system. For personalized services, a successful recommendation consists of two parts: attracting users to click the item and users being willing to consume the item. If both tasks need to be predicted at the same time, traditional recommendation systems generally train two independent models. This approach is cumbersome and does not effectively model the relationship between the two subtasks of "click-consumption". Therefore, in order to improve the success rate of recommendation and reduce computational costs, researchers are trying to model multi-task learning. At present, existing multi-task learning models generally adopt hard parameter sharing or soft parameter sharing architecture, but these two architectures each have certain problems. Therefore, in this work, we propose a novel recommendation model based on real recommendation scenarios, Deep Cross network based on RNN for partial parameter sharing (DCRNN). The model has three innovations: 1) It adopts the idea of cross network and uses RNN network to cross-process the features, thereby effectively improves the expressive ability of the model; 2) It innovatively proposes the structure of partial parameter sharing; 3) It can effectively capture the potential correlation between different tasks to optimize the efficiency and methods for learning different tasks.
Underwater Sound Speed Profile (SSP) distribution has great influence on the propagation mode of acoustic signal, thus the fast and accurate estimation of SSP is of great importance in building underwater observation systems. The state-of-the-art SSP inversion methods include frameworks of matched field processing (MFP), compressive sensing (CS), and feedforeward neural networks (FNN), among which the FNN shows better real-time performance while maintain the same level of accuracy. However, the training of FNN needs quite a lot historical SSP samples, which is diffcult to be satisfied in many ocean areas. This situation is called few-shot learning. To tackle this issue, we propose a multi-task learning (MTL) model with partial parameter sharing among different traning tasks. By MTL, common features could be extracted, thus accelerating the learning process on given tasks, and reducing the demand for reference samples, so as to enhance the generalization ability in few-shot learning. To verify the feasibility and effectiveness of MTL, a deep-ocean experiment was held in April 2023 at the South China Sea. Results shows that MTL outperforms the state-of-the-art methods in terms of accuracy for SSP inversion, while inherits the real-time advantage of FNN during the inversion stage.
Characterizing multipartite quantum systems is crucial for quantum computing and many-body physics. The problem, however, becomes challenging when the system size is large and the properties of interest involve correlations among a large number of particles. Here we introduce a neural network model that can predict various quantum properties of many-body quantum states with constant correlation length, using only measurement data from a small number of neighboring sites. The model is based on the technique of multi-task learning, which we show to offer several advantages over traditional single-task approaches. Through numerical experiments, we show that multi-task learning can be applied to sufficiently regular states to predict global properties, like string order parameters, from the observation of short-range correlations, and to distinguish between quantum phases that cannot be distinguished by single-task networks. Remarkably, our model appears to be able to transfer information learnt from lower dimensional quantum systems to higher dimensional ones, and to make accurate predictions for Hamiltonians that were not seen in the training.
Parameter-efficient fine-tuning (PEFT) has shown its effectiveness in adapting the pre-trained language models to downstream tasks while only updating a small number of parameters. Despite the success, most existing methods independently adapt to each task without considering knowledge transfer between tasks and are limited to low-data regimes. To overcome this issue, we propose Prototype-based HyperAdapter (PHA), a novel framework built on the adapter-tuning and hypernetwork. It introduces an instance-dense retriever and a prototypical hypernetwork to generate the conditional modules in a sample-efficient manner. This leads to comparable performance improvements against existing PEFT methods on multi-task learning and few-shot transfer learning. More importantly, when the available data size gets smaller, our method outperforms other strong baselines by a large margin. Based on our extensive empirical experiments across various datasets, we demonstrate that PHA strikes a better trade-off between trainable parameters, accuracy on stream tasks, and sample efficiency.