CVPR 2020 continual learning in computer vision competition: Approaches, results, current challenges and future directions

Lomonaco, Vincenzo; Pellegrini, Lorenzo; Rodríguez, Paul; Caccia, M.; She, Qi; Yu, Chen; Jodelet, Quentin; Wang, Ruiping; Mai, Zheda; Vázquez, David; Parisi, German I.; Churamani, Nikhil; Pickett, Marc; Laradji, Issam; Maltoni, Davide

doi:10.1016/j.artint.2021.103635

Cited by 33 publications

(16 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Rehearsal-based methods, i.e., approaches that leverage a memory buffer to cope with catastrophic forgetting, are emerging as the most effective methodology to tackle CL. Their performance, backed by extensive empirical evidence (16), finds also a theoretical justification in Knoblauch and coworkers' finding that optimally solving CL would require perfect memory of the past (17). In fact, if we were able to completely re-train a new system with all previous data every time a new task arrives, Continual Learning would not appear to be any different from any other learning problem.…”

Section: Introductionmentioning

confidence: 95%

“…Solutions to this problem typically incur in a increase in resource requirements (16) both for CL's very nature (the more tasks arrive the more data the agent need to process), and for the nature of the systems that try to solve it, both in the increased complexity of the typically deep learning models, and in the time and space requirements of continuously learning multiple models. This problem become particularly evident in rehearsal-based methods.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Smaller Is Better: An Analysis of Instance Quantity/Quality Trade-off in Rehearsal-based Continual Learning

Pelosin¹,

Torsello²

2021

Preprint

View full text Add to dashboard Cite

The design of machines and algorithms capable of learning in a dynamically changing environment has become an increasingly topical problem with the increase of the size and heterogeneity of data available to learning systems. As a consequence, the key issue of Continual Learning has become that of addressing the stabilityplasticity dilemma of connectionist systems, as they need to adapt their model without forgetting previously acquired knowledge. Within this context, rehearsalbased methods i.e., solutions in where the learner exploits memory to revisit past data, has proven to be very effective, leading to performance at the state-of-the-art. In our study, we propose an analysis of the memory quantity/quality trade-off adopting various data reduction approaches to increase the number of instances storable in memory. In particular, we investigate complex instance compression techniques such as deep encoders, but also trivial approaches such as image resizing and linear dimensionality reduction. Our findings suggest that the optimal trade-off is severely skewed toward instance quantity, where rehearsal approaches with several heavily compressed instances easily outperform state-of-the-art approaches with the same amount of memory at their disposal. Further, in high memory configurations, deep approaches extracting spatial structure combined with extreme resizing (of the order of 8 × 8 images) yield the best results, while in memory-constrained configurations where deep approaches cannot be used due to their memory requirement in training, Extreme Learning Machines (ELM) offer a clear advantage.Preprint. Under review.

show abstract

Section: Introductionmentioning

confidence: 95%

Section: Introductionmentioning

confidence: 99%

Smaller Is Better: An Analysis of Instance Quantity/Quality Trade-off in Rehearsal-based Continual Learning

Pelosin¹,

Torsello²

2021

Preprint

View full text Add to dashboard Cite

show abstract

“…This is continual learning (Chen and Liu 2018), which has been discussed earlier. We will not discuss it further as there are already numerous existing techniques (Parisi et al 2019;Lomonaco et al 2022). Many can leverage existing knowledge to learn the new task better (Chen and Liu 2018).…”

Section: Open World Continual Learningmentioning

confidence: 99%

“…Although novelty detection (Yang et al 2021;Pang et al 2021) and incremental or continual learning (Chen and Liu 2018; Parisi et al 2019;Lomonaco et al 2022) have been studied widely, little work has been done to build a SOLA system. Here we describe a dialogue system (called CML) that is based on the SOLA framework and performs each function in SOLA continually by itself on the job during conversation (Mazumder et al 2020b).…”

Section: Cml: An Example Sola Systemmentioning

confidence: 99%

AI Autonomy: Self-Initiation, Adaptation and Continual Learning

Liu¹,

Mazumder²,

Robertson³

et al. 2022

Preprint

View full text Add to dashboard Cite

As more and more AI agents are used in practice, it is time to think about how to make these agents fully autonomous so that they can (1) learn by themselves continually in a selfmotivated and self-initiated manner rather than being retrained offline periodically on the initiation of human engineers and (2) accommodate or adapt to unexpected or novel circumstances. As the real-world is an open environment that is full of unknowns or novelties, detecting novelties, characterizing them, accommodating or adapting to them, and gathering ground-truth training data and incrementally learning the unknowns/novelties are critical to making the AI agent more and more knowledgeable and powerful over time. The key challenge is how to automate the process so that it is carried out continually on the agent's own initiative and through its own interactions with humans, other agents and the environment just like human on-the-job learning. This paper proposes a framework (called SOLA) for this learning paradigm to promote the research of building autonomous and continual learning enabled AI agents. To show feasibility, an implemented agent is also described.

show abstract

“…The rapid growth of continual learning has lead researchers to work on empirical studies [20,58,56], surveys [32,42,55,66,67] as well as CL-specific software [72,22,59].…”

Section: Related Workmentioning

confidence: 99%

Continual Learning via Local Module Composition

Ostapenko¹,

Rodríguez²,

Caccia³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

Modularity is a compelling solution to continual learning (CL), the problem of modeling sequences of related tasks. Learning and then composing modules to solve different tasks provides an abstraction to address the principal challenges of CL including catastrophic forgetting, backward and forward transfer across tasks, and sub-linear model growth. We introduce local module composition (LMC), an approach to modular CL where each module is provided a local structural component that estimates a module's relevance to the input. Dynamic module composition is performed layer-wise based on local relevance scores. We demonstrate that agnosticity to task identities (IDs) arises from (local) structural learning that is module-specific as opposed to the task-and/or model-specific as in previous works, making LMC applicable to more CL settings compared to previous works. In addition, LMC also tracks statistics about the input distribution and adds new modules when outlier samples are detected. In the first set of experiments, LMC performs favorably compared to existing methods on the recent Continual Transfer-learning Benchmark without requiring task identities. In another study, we show that the locality of structural learning allows LMC to interpolate to related but unseen tasks (OOD), as well as to compose modular networks trained independently on different task sequences into a third modular network without any fine-tuning. Finally, in search for limitations of LMC we study it on more challenging sequences of 30 and 100 tasks, demonstrating that local module selection becomes much more challenging in presence of a large number of candidate modules. In this setting best performing LMC spawns much fewer modules compared to an oracle based baseline, however it reaches a lower overall accuracy. The codebase is available under https://github.com/oleksost/LMC.

show abstract

CVPR 2020 continual learning in computer vision competition: Approaches, results, current challenges and future directions

Cited by 33 publications

References 12 publications

Smaller Is Better: An Analysis of Instance Quantity/Quality Trade-off in Rehearsal-based Continual Learning

Smaller Is Better: An Analysis of Instance Quantity/Quality Trade-off in Rehearsal-based Continual Learning

AI Autonomy: Self-Initiation, Adaptation and Continual Learning

Continual Learning via Local Module Composition

Contact Info

Product

Resources

About