Classes, as fundamental elements of Computer Vision, have been extensively studied within incremental learning frameworks. In contrast, tokens, which play essential roles in many research fields, exhibit similar characteristics of growth, yet investigations into their incremental learning remain significantly scarce. This research gap primarily stems from the holistic nature of tokens in language, which imposes significant challenges on the design of incremental learning frameworks for them. To overcome this obstacle, in this work, we turn to a type of token, gene, for a large-scale biological dataset--single-cell transcriptomics--to formulate a pipeline for gene incremental learning and establish corresponding evaluations. We found that the forgetting problem also exists in gene incremental learning, thus we adapted existing class incremental learning methods to mitigate the forgetting of genes. Through extensive experiments, we demonstrated the soundness of our framework design and evaluations, as well as the effectiveness of our method adaptations. Finally, we provide a complete benchmark for gene incremental learning in single-cell transcriptomics.
We derive distance relay characteristics in terms of incremental phasors. We use a circuit model of the network to estimate the incremental remote current. If we assume that all sources are stationary, i.e., remain periodic shortly after a fault, then the incremental remote current, and thus the characteristics, do not depend on the real-time voltages or current injections of the sources.
Real-world graphs are dynamic, with frequent updates to their structure and features due to evolving vertex and edge properties. These continual changes pose significant challenges for efficient inference in graph neural networks (GNNs). Existing vertex-wise and layer-wise inference approaches are ill-suited for dynamic graphs, as they incur redundant computations, large neighborhood traversals, and high communication costs, especially in distributed settings. Additionally, while sampling-based approaches can be adopted to approximate final layer embeddings, these are often not preferred in critical applications due to their non-determinism. These limitations hinder low-latency inference required in real-time applications. To address this, we propose RIPPLE++, a framework for streaming GNN inference that efficiently and accurately updates embeddings in response to changes in the graph structure or features. RIPPLE++ introduces a generalized incremental programming model that captures the semantics of GNN aggregation functions and incrementally propagates updates to affected neighborhoods. RIPPLE++ accommodates all common graph updates, including vertex/edge addition/deletions and v
A theorem is proved to verify incremental stability of a feedback system via a homotopy from a known incrementally stable system. A first corollary of that result is that incremental stability may be verified by separation of Scaled Relative Graphs, correcting two assumptions in [1, Theorem 2]. A second corollary provides an incremental version of the classical IQC stability theorem.
We study incremental constituent parsers to assess their capacity to output trees based on prefix representations alone. Guided by strictly left-to-right generative language models and tree-decoding modules, we build parsers that adhere to a strong definition of incrementality across languages. This builds upon work that asserted incrementality, but that mostly only enforced it on either the encoder or the decoder. Finally, we conduct an analysis against non-incremental and partially incremental models.
Incremental few-shot object detection aims at detecting novel classes without forgetting knowledge of the base classes with only a few labeled training data from the novel classes. Most related prior works are on incremental object detection that rely on the availability of abundant training samples per novel class that substantially limits the scalability to real-world setting where novel data can be scarce. In this paper, we propose the Incremental-DETR that does incremental few-shot object detection via fine-tuning and self-supervised learning on the DETR object detector. To alleviate severe over-fitting with few novel class data, we first fine-tune the class-specific components of DETR with self-supervision from additional object proposals generated using Selective Search as pseudo labels. We further introduce an incremental few-shot fine-tuning strategy with knowledge distillation on the class-specific components of DETR to encourage the network in detecting novel classes without forgetting the base classes. Extensive experiments conducted on standard incremental object detection and incremental few-shot object detection settings show that our approach significantly outperform
Incremental methods are widely utilized for solving finite-sum optimization problems in machine learning and signal processing. In this paper, we study a family of incremental methods -- including incremental subgradient, incremental proximal point, and incremental prox-linear methods -- for solving weakly convex optimization problems. Such a problem class covers many nonsmooth nonconvex instances that arise in engineering fields. We show that the three said incremental methods have an iteration complexity of $O(\varepsilon^{-4})$ for driving a natural stationarity measure to below $\varepsilon$. Moreover, we show that if the weakly convex function satisfies a sharpness condition, then all three incremental methods, when properly initialized and equipped with geometrically diminishing stepsizes, can achieve a local linear rate of convergence. Our work is the first to extend the convergence rate analysis of incremental methods from the nonsmooth convex regime to the weakly convex regime. Lastly, we conduct numerical experiments on the robust matrix sensing problem to illustrate the convergence performance of the three incremental methods.
Property Directed Reachability (PDR) is a widely used technique for formal verification of hardware and software systems. This paper presents an incremental version of PDR (IPDR), which enables the automatic verification of system instances of incremental complexity. The proposed algorithm leverages the concept of incremental SAT solvers to reuse verification results from previously verified system instances, thereby accelerating the verification process. The new algorithm supports both incremental constraining and relaxing; i.e., starting from an over-constrained instance that is gradually relaxed. To validate the effectiveness of the proposed algorithm, we implemented IPDR and experimentally evaluate it on two different problem domains. First, we consider a circuit pebbling problem, where the number of pebbles is both constrained and relaxed. Second, we explore parallel program instances, progressively increasing the allowed number of interleavings. The experimental results demonstrate significant performance improvements compared to Z3's PDR implementation SPACER. Experiments also show that the incremental approach succeeds in reusing a substantial amount of clauses between inst
Humans are capable of acquiring new knowledge and transferring learned knowledge into different domains, incurring a small forgetting. The same ability, called Continual Learning, is challenging to achieve when operating with neural networks due to the forgetting affecting past learned tasks when learning new ones. This forgetting can be mitigated by replaying stored samples from past tasks, but a large memory size may be needed for long sequences of tasks; moreover, this could lead to overfitting on saved samples. In this paper, we propose a novel regularisation approach and a novel incremental classifier called, respectively, Margin Dampening and Cascaded Scaling Classifier. The first combines a soft constraint and a knowledge distillation approach to preserve past learned knowledge while allowing the model to learn new patterns effectively. The latter is a gated incremental classifier, helping the model modify past predictions without directly interfering with them. This is achieved by modifying the output of the model with auxiliary scaling functions. We empirically show that our approach performs well on multiple benchmarks against well-established baselines, and we also study
Incremental computation aims to compute more efficiently on changed input by reusing previously computed results. We give a high-level overview of works on incremental computation, and highlight the essence underlying all of them, which we call incrementalization -- the discrete counterpart of differentiation in calculus. We present the gist of a systematic method for incrementalization, and a systematic method centered around it -- called Iterate-Incrementalize-Implement -- for program design and optimization, as well as algorithm design and optimization. We illustrate the methods with example applications in arithmetic computations, recursive functions, graph analysis, and distributed algorithms. At a meta-level, with historical contexts and for future directions, we stress the power of high-level data, control, and module abstractions in developing new and better algorithms and programs as well as their precise complexities.
Many explainable AI (XAI) techniques strive for interpretability by providing concise salient information, such as sparse linear factors. However, users either only see inaccurate global explanations, or highly-varying local explanations. We propose to provide more detailed explanations by leveraging the human cognitive capacity to accumulate knowledge by incrementally receiving more details. Focusing on linear factor explanations (factors $\times$ values = outcome), we introduce Incremental XAI to automatically partition explanations for general and atypical instances by providing Base + Incremental factors to help users read and remember more faithful explanations. Memorability is improved by reusing base factors and reducing the number of factors shown in atypical cases. In modeling, formative, and summative user studies, we evaluated the faithfulness, memorability and understandability of Incremental XAI against baseline explanation methods. This work contributes towards more usable explanation that users can better ingrain to facilitate intuitive engagement with AI.
Incremental SAT and QBF solving potentially yields improvements when sequences of related formulas are solved. An incremental application is usually tailored towards some specific solver and decomposes a problem into incremental solver calls. This hinders the independent comparison of different solvers, particularly when the application program is not available. As a remedy, we present an approach to automated benchmarking of incremental SAT and QBF solvers. Given a collection of formulas in (Q)DIMACS format generated incrementally by an application program, our approach automatically translates the formulas into instructions to import and solve a formula by an incremental SAT/QBF solver. The result of the translation is a program which replays the incremental solver calls and thus allows to evaluate incremental solvers independently from the application program. We illustrate our approach by different hardware verification problems for SAT and QBF solvers.
Incremental processing allows interactive systems to respond based on partial inputs, which is a desirable property e.g. in dialogue agents. The currently popular Transformer architecture inherently processes sequences as a whole, abstracting away the notion of time. Recent work attempts to apply Transformers incrementally via restart-incrementality by repeatedly feeding, to an unchanged model, increasingly longer input prefixes to produce partial outputs. However, this approach is computationally costly and does not scale efficiently for long sequences. In parallel, we witness efforts to make Transformers more efficient, e.g. the Linear Transformer (LT) with a recurrence mechanism. In this work, we examine the feasibility of LT for incremental NLU in English. Our results show that the recurrent LT model has better incremental performance and faster inference speed compared to the standard Transformer and LT with restart-incrementality, at the cost of part of the non-incremental (full sequence) quality. We show that the performance drop can be mitigated by training the model to wait for right context before committing to an output and that training with input prefixes is beneficial
Data is continuously generated by modern data sources, and a recent challenge in machine learning has been to develop techniques that perform well in an incremental (streaming) setting. In this paper, we investigate the problem of private machine learning, where as common in practice, the data is not given at once, but rather arrives incrementally over time. We introduce the problems of private incremental ERM and private incremental regression where the general goal is to always maintain a good empirical risk minimizer for the history observed under differential privacy. Our first contribution is a generic transformation of private batch ERM mechanisms into private incremental ERM mechanisms, based on a simple idea of invoking the private batch ERM procedure at some regular time intervals. We take this construction as a baseline for comparison. We then provide two mechanisms for the private incremental regression problem. Our first mechanism is based on privately constructing a noisy incremental gradient function, which is then used in a modified projected gradient procedure at every timestep. This mechanism has an excess empirical risk of $\approx\sqrt{d}$, where $d$ is the dimen
While humans process language incrementally, the best language encoders currently used in NLP do not. Both bidirectional LSTMs and Transformers assume that the sequence that is to be encoded is available in full, to be processed either forwards and backwards (BiLSTMs) or as a whole (Transformers). We investigate how they behave under incremental interfaces, when partial output must be provided based on partial input seen up to a certain time step, which may happen in interactive systems. We test five models on various NLU datasets and compare their performance using three incremental evaluation metrics. The results support the possibility of using bidirectional encoders in incremental mode while retaining most of their non-incremental quality. The "omni-directional" BERT model, which achieves better non-incremental performance, is impacted more by the incremental access. This can be alleviated by adapting the training regime (truncated training), or the testing procedure, by delaying the output until some right context is available or by incorporating hypothetical right contexts generated by a language model like GPT-2.
Deep learning research over the past years has shown that by increasing the scope or difficulty of the learning problem over time, increasingly complex learning problems can be addressed. We study incremental learning in the context of sequence learning, using generative RNNs in the form of multi-layer recurrent Mixture Density Networks. While the potential of incremental or curriculum learning to enhance learning is known, indiscriminate application of the principle does not necessarily lead to improvement, and it is essential therefore to know which forms of incremental or curriculum learning have a positive effect. This research contributes to that aim by comparing three instantiations of incremental or curriculum learning. We introduce Incremental Sequence Learning, a simple incremental approach to sequence learning. Incremental Sequence Learning starts out by using only the first few steps of each sequence as training data. Each time a performance criterion has been reached, the length of the parts of the sequences used for training is increased. We introduce and make available a novel sequence learning task and data set: predicting and classifying MNIST pen stroke sequences.
We introduce incremental variational inference and apply it to latent Dirichlet allocation (LDA). Incremental variational inference is inspired by incremental EM and provides an alternative to stochastic variational inference. Incremental LDA can process massive document collections, does not require to set a learning rate, converges faster to a local optimum of the variational bound and enjoys the attractive property of monotonically increasing it. We study the performance of incremental LDA on large benchmark data sets. We further introduce a stochastic approximation of incremental variational inference which extends to the asynchronous distributed setting. The resulting distributed algorithm achieves comparable performance as single host incremental variational inference, but with a significant speed-up.
Blockwise self-attentional encoder models have recently emerged as one promising end-to-end approach to simultaneous speech translation. These models employ a blockwise beam search with hypothesis reliability scoring to determine when to wait for more input speech before translating further. However, this method maintains multiple hypotheses until the entire speech input is consumed -- this scheme cannot directly show a single \textit{incremental} translation to users. Further, this method lacks mechanisms for \textit{controlling} the quality vs. latency tradeoff. We propose a modified incremental blockwise beam search incorporating local agreement or hold-$n$ policies for quality-latency control. We apply our framework to models trained for online or offline translation and demonstrate that both types can be effectively used in online mode. Experimental results on MuST-C show 0.6-3.6 BLEU improvement without changing latency or 0.8-1.4 s latency improvement without changing quality.
In this paper, we consider an adaptive approach to address optimization problems with uncertain cost parameters. Here, the decision maker selects an initial decision, observes the realization of the uncertain cost parameters, and then is permitted to modify the initial decision. We treat the uncertainty using the framework of robust optimization in which uncertain parameters lie within a given set. The decision maker optimizes so as to develop the best cost guarantee in terms of the worst-case analysis. The recourse decision is ``incremental"; that is, the decision maker is permitted to change the initial solution by a small fixed amount. We refer to the resulting problem as the robust incremental problem. We study robust incremental variants of several optimization problems. We show that the robust incremental counterpart of a linear program is itself a linear program if the uncertainty set is polyhedral. Hence, it is solvable in polynomial time. We establish the NP-hardness for robust incremental linear programming for the case of a discrete uncertainty set. We show that the robust incremental shortest path problem is NP-complete when costs are chosen from a polyhedral uncertaint
Multi-class supervised learning systems require the knowledge of the entire range of labels they predict. Often when learnt incrementally, they suffer from catastrophic forgetting. To avoid this, generous leeways have to be made to the philosophy of incremental learning that either forces a part of the machine to not learn, or to retrain the machine again with a selection of the historic data. While these hacks work to various degrees, they do not adhere to the spirit of incremental learning. In this article, we redefine incremental learning with stringent conditions that do not allow for any undesirable relaxations and assumptions. We design a strategy involving generative models and the distillation of dark knowledge as a means of hallucinating data along with appropriate targets from past distributions. We call this technique, phantom sampling.We show that phantom sampling helps avoid catastrophic forgetting during incremental learning. Using an implementation based on deep neural networks, we demonstrate that phantom sampling dramatically avoids catastrophic forgetting. We apply these strategies to competitive multi-class incremental learning of deep neural networks. Using vari