CBR Meets Big Data: A Case Study of Large-Scale Adaptation Rule Generation

Jalali, Vahid; Leake, David

doi:10.1007/978-3-319-24586-7_13

Cited by 10 publications

(5 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…the sample presented in section 4), 13 are modeled in the strategies above: 1 in §6.1, 1 in §6.2, 9 in §6.3 and 2 in §6.4 7 (0 in §6.5). 5 of the remaining ones corresponds to a strategy consisting in adding or substituting a qualifier to the source recipe name when a new ingredient is added or replaces an ingredient that has no connection with the source recipe name.…”

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

Analogical Transfer in RDFS, Application to Cocktail Name Adaptation

Kiani¹,

Lieber²,

Nauer³

et al. 2016

Case-Based Reasoning Research and Development

View full text Add to dashboard Cite

Abstract. This paper deals with analogical transfer in the framework of the representation language RDFS. The application of analogical transfer to case-based reasoning consists in reusing the problem-solution dependency to the context of the target problem; thus it is a general approach to adaptation. RDFS is a representation language that is a standard of the semantic Web; it is based on RDF, a graphical representation of data, completed by an entailment relation. A dependency is therefore represented as a graph representing complex links between a problem and a solution, and analogical transfer uses, in particular, RDFS entailment. This research work is applied (and inspired from) the issue of cocktail name adaptation: given a cocktail and a way this cocktail is adapted by changing its ingredient list, how can the cocktail name be modified?

show abstract

Section: Discussionmentioning

confidence: 99%

“…Adaptation is a research issue of case-based reasoning (CBR [11]) that has received some attention during the last years in the CBR community (see, e.g., [2,7,9]). In particular, this has been an issue for the competitors of the Computer Cooking Contests (CCCs).…”

Section: Introductionmentioning

confidence: 99%

Analogical Transfer in RDFS, Application to Cocktail Name Adaptation

Kiani¹,

Lieber²,

Nauer³

et al. 2016

Case-Based Reasoning Research and Development

View full text Add to dashboard Cite

show abstract

“…LSH maps similar datapoints (cases) into same 'buckets' (encoded with low-dimensional binary codes) with high probability, which preserves the similarity relationships between datapoints and can be regarded as a way of dimensionality reduction on high-dimensional data [14]. Several studies introduce LSH into CBR systems to approximate the nearest neighbor search process and scale traditional CBR systems to large-scale data [29,30,70]. The studies show CBR systems equipped with hashing techniques can greatly improve retrieval efficiency and achieve desirable performance with expected loss in accuracy.…”

Section: Gaps Of Efficient Case-based Reasoningmentioning

confidence: 99%

Supervised Deep Hashing for High-dimensional and Heterogeneous Case-based Reasoning

Zhang¹,

Liang²,

Shi³

et al. 2022

Preprint

View full text Add to dashboard Cite

Case-based Reasoning (CBR) on high-dimensional and heterogeneous data is a trending yet challenging and computationally expensive task in the real world. A promising approach is to obtain low-dimensional hash codes representing cases and perform a similarity retrieval of cases in a Hamming space. However, previous methods based on data-independent hashing rely on random projections or manual construction, inapplicable to address specific data issues (e.g., high-dimensionality and heterogeneity) due to their insensitivity to data characteristics. To address these issues, this work introduces a novel deep hashing network to learn similarity-preserving compact hash codes for efficient case retrieval and proposes a deep-hashing-enabled CBR model HeCBR. Specifically, we introduce position embedding to represent heterogeneous features and utilize a multilinear interaction layer to obtain case embeddings, which effectively filtrates zero-valued features to tackle high-dimensionality and sparsity and captures inter-feature couplings. Then, we feed the case embeddings into fully-connected layers, and subsequently a hash layer generates hash codes with a quantization regularizer to control the quantization loss during relaxation. To cater for incremental learning of CBR, we further propose an adaptive learning strategy to update the hash function. Extensive experiments on public datasets show HeCBR greatly reduces storage and significantly accelerates the case retrieval. HeCBR achieves desirable performance compared with the state-of-the-art CBR methods and performs significantly better than hashing-based CBR methods in classification.

show abstract

“…[9] introduces this approach of AK learning based on pairwise comparisons of cases. This approach, also called Case Difference Heuristic in [10], has been applied in various domains such as medicine [3] or cooking [1,5]. So, for ordered pairs of cases (c , c r ) associated to their variations V r forming a formal context, as the one presented in Table 2, an AK discovery process based on FCI can be run.…”

Section: Exploiting Case Variations For Adaptation Knowledge Discovery With Positive and Negative Casesmentioning

confidence: 99%

Adaptation Knowledge Discovery Using Positive and Negative Cases

Lieber

Nauer

2021

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Case-based reasoning usually exploits positive source cases, each of them consisting in a problem and a correct solution to this problem. Now, the general issue of exploiting also negative cases-i.e., problem-solution pairs where the solution answers incorrectly the problem-can be raised. Indeed, such cases are "naturally" generated by a CBR system as long as it sometimes proposes incorrect solutions. This paper aims at addressing this issue for adaptation knowledge (AK) discovery: how positive and negative cases can be used for this purpose. The idea is that positive cases are used to propose adaptation rules and that negative cases are used to filter out some of these rules. In a preliminary work, this kind of AK discovery has been applied using frequent closed itemset (FCI) extraction on variations within the case base and tested on a toy Boolean use case, with promising first results. This paper resumes this study and evaluates it on 4 benchmarks, which confirms the benefit of exploiting negative cases for AK discovery. This involves some adjustments in the data preparation and in adaptation rule filtering, in particular because FCI extraction works only with Boolean features, hence some methodology lessons learned for AK discovery with positive and negative cases.

show abstract

CBR Meets Big Data: A Case Study of Large-Scale Adaptation Rule Generation

Cited by 10 publications

References 25 publications

Analogical Transfer in RDFS, Application to Cocktail Name Adaptation

Analogical Transfer in RDFS, Application to Cocktail Name Adaptation

Supervised Deep Hashing for High-dimensional and Heterogeneous Case-based Reasoning

Adaptation Knowledge Discovery Using Positive and Negative Cases

Contact Info

Product

Resources

About