Filters








5,785 Hits in 7.0 sec

Let the CAT out of the bag: Contrastive Attributed explanations for Text [article]

Saneem Chemmengath, Amar Prakash Azad, Ronny Luss, Amit Dhurandhar
2021 arXiv   pre-print
In this paper, we propose a method Contrastive Attributed explanations for Text (CAT) which provides contrastive explanations for natural language text data with a novel twist as we build and exploit attribute  ...  Contrastive explanations for understanding the behavior of black box models has gained a lot of attention recently as they provide potential for recourse.  ...  We demonstrate the utility of CAT through a user study that asks users to determine a model's prediction on an instance given an explanation.  ... 
arXiv:2109.07983v1 fatcat:6ltgdbtmtbhqjbv23f6gcxinvu

Swap Retrieval

Amir Ghodrati, Xu Jia, Marco Pedersoli, Tinne Tuytelaars
2015 Proceedings of the 5th ACM on International Conference on Multimedia Retrieval - ICMR '15  
For instance, starting from an image of a dog in a certain situation/context, the goal is to find images of cats with a similar situation/context.  ...  In this work we make a first step towards getting the best of both worlds: we use an image to specify the context, but let the user specify a related category as main search criterion.  ...  In this respect, letting the method figure out what are the important characteristics of an image and retrieve similar ones can give more satisfactory results than starting from a text-based query.  ... 
doi:10.1145/2671188.2749373 dblp:conf/mir/GhodratiJPT15 fatcat:6lj7bicgyvaxbj6b726vwusg2u

Contrastive knowledge

Antti Karjalainen, Adam Morton
2003 Philosophical Explorations  
Neither Lipton nor we would intend this to be precise, let alone a definition of contrastive explanation.  ...  One way of capturing this fact is to let the knowledge ascription inherit the contrastivity of the evidence.  ... 
doi:10.1080/10002003058538741 fatcat:iwpw67yfxjcc7a6kcdwen6v3gm

Semantic Supervision: Enabling Generalization over Output Spaces [article]

Austin W. Hanjie and Ameet Deshpande and Karthik Narasimhan
2022 arXiv   pre-print
In contrast to standard classification, which treats classes as discrete symbols, SemSup represents them as dense vector features obtained from descriptions of classes (e.g., "The cat is a small carnivorous  ...  This allows the output space to be unbounded (in the space of descriptions) and enables models to generalize both over unseen inputs and unseen outputs (e.g.  ...  Acknowledgements We are grateful to Tianyu Gao, Shunyu Yao, Shreyas Chaudhari (UMass Amherst), and members of the Princeton NLP group for their feedback and suggestions.  ... 
arXiv:2202.13100v2 fatcat:ix46aq7affdgrkwcdxqp62dfsa

John McGahern's stylistic and narratological art

Beatrix Busse, Dan McIntyre, Nina Nørgaard, Michael Toolan
2010 Lexis: Journal in English Lexicology  
The cat, for example, is thrust into a canvas bag and drowned in water.  ...  We speculate that (McGahern conceived that) the couple's routine would be for Agnes to bring the Tesco bags of shopping out of the supermarket, their contents then to be transferred to Michael's discreet  ... 
doi:10.4000/lexis.530 fatcat:4zonh7j6ojeg5ctxwjnvplymxi

How does this interaction affect me? Interpretable attribution for feature interactions [article]

Michael Tsang, Sirisha Rambhatla, Yan Liu
2020 arXiv   pre-print
Machine learning transparency calls for interpretable explanations of how inputs relate to predictions. Feature attribution is a way to analyze the impact of features on predictions.  ...  Our experiments on standard annotation labels indicate our approach provides significantly more interpretable explanations than comparable methods, which is important for analyzing the impact of interactions  ...  Figure 7 : 7 Figure 7: Text explanation metrics ((a) Word ρ and (b) Phrase ρ) versus top and bottom % of attributions retained for different attribution methods on BERT over the SST test set.  ... 
arXiv:2006.10965v1 fatcat:xh2kexnhkjemni5tj2mblkeucq

Linked knowledge sources for topic classification of microposts: A semantic graph-based approach

Andrea Varga, Amparo Elizabeth Cano Basave, Matthew Rowe, Fabio Ciravegna, Yulan He
2014 Journal of Web Semantics  
For this reason we propose an approach for automatic estimation of accuracy loss of a topic classifier on new, unseen Microposts.  ...  In contrast our approach relies on the generation of different weighted semantic meta-graphs extracted from linked KSs. We introduce a new semantic graph, called category meta-graph.  ...  Acknowledgement The authors would like to express their gratitude to Aba-Sah Dadzie for proofreading this paper.  ... 
doi:10.1016/j.websem.2014.04.001 fatcat:64qvdi3snramrovjqhxuimcpwe

Linked Knowledge Sources for Topic Classification of Microposts: A Semantic Graph-Based Approach

Andrea Varga, Amparo E. Cano, Matthew Rowe, Fabio Ciravegna, Yulan He
2014 Social Science Research Network  
For this reason we propose an approach for automatic estimation of accuracy loss of a topic classifier on new, unseen Microposts.  ...  In contrast our approach relies on the generation of different weighted semantic meta-graphs extracted from linked KSs. We introduce a new semantic graph, called category meta-graph.  ...  Acknowledgement The authors would like to express their gratitude to Aba-Sah Dadzie for proofreading this paper.  ... 
doi:10.2139/ssrn.3199099 fatcat:2febxkwyfbc6jisntu24z4ro3a

Benchmarking and Survey of Explanation Methods for Black Box Models [article]

Francesco Bodria, Fosca Giannotti, Riccardo Guidotti, Francesca Naretto, Dino Pedreschi, Salvatore Rinzivillo
2021 arXiv   pre-print
The widespread adoption of black-box models in Artificial Intelligence has enhanced the need for explanation methods to reveal how these obscure models reach specific decisions.  ...  We provide a categorization of explanation methods based on the type of explanation returned.  ...  For the experiments, we adopt the same preprocessing used for lime and intgrad. L2X [29] can produce a SH explanation for text. In particular, for text, the patches are now a group of words.  ... 
arXiv:2102.13076v1 fatcat:6iw6nhwltzb7fhaawnvy7oxndu

The Role of Common-Sense Knowledge in Assessing Semantic Association

Alice Ruggeri, Luigi Di Caro, Guido Boella
2018 Journal on Data Semantics  
explanation (and the use) of relatedness and similarity with non-numeric information.  ...  Natural language processing techniques often aim at automatically extracting semantics from texts.  ...  For instance, the analysis of the words bear and cat may lead to explanations that associate different f unctional features to the property of having claws (e.g., climbing trees for a cat, killing people  ... 
doi:10.1007/s13740-018-0094-2 fatcat:6e4exrmrdvby7gnayo4cxkgl6q

Domain-Specific Image Captioning

Rebecca Mason, Eugene Charniak
2014 Proceedings of the Eighteenth Conference on Computational Natural Language Learning  
We extract previously-written descriptions from a database and adapt them to new query images, using a joint visual and textual bag-of-words model to determine the correctness of individual words.  ...  We present a data-driven framework for image caption generation which incorporates visual and textual features with varying degrees of spatial structure.  ...  z img and z txt are drawn for each of the text words and codewords: z img ∼ P (z img |θ) = n θ z img n (2) 6 While space limits a more detailed explanation of visual bag-of-word features, Section 5.2  ... 
doi:10.3115/v1/w14-1602 dblp:conf/conll/MasonC14 fatcat:yylrtfalabfohie2ifp5qc7qje

Document re-ranking based on topic-comment structure

Liana Ermakova, Josiane Mothe
2016 2016 IEEE Tenth International Conference on Research Challenges in Information Science (RCIS)  
analysis of discourse to capture document relevance by distinguishing the topic of a text from what is said about the topic (comment) in the text.  ...  While most information retrieval models make the assumption that relevant documents are about the query and that aboutness can be captured considering bags of words only, we rather consider a more sophisticated  ...  Acknowledgments: The authors would like to thank Ambassade de France en Russie for funding this project (bourse de thèse en cotutelle).  ... 
doi:10.1109/rcis.2016.7549352 dblp:conf/rcis/ErmakovaM16 fatcat:6ksxfh5nzfafnhs432yr4x3xkq

Wikipedia-based Semantic Interpretation for Natural Language Processing

E. Gabrilovich, S. Markovitch
2009 The Journal of Artificial Intelligence Research  
We evaluate the effectiveness of our method on text categorization and on computing the degree of semantic relatedness between fragments of natural language text.  ...  Here we propose a novel method, called Explicit Semantic Analysis (ESA), for fine-grained semantic interpretation of unrestricted natural language texts.  ...  This work was partially supported by funding from the EC-sponsored MUSCLE Network of Excellence. The first author's current address is Yahoo!  ... 
doi:10.1613/jair.2669 fatcat:mwcky2jqx5e6zimhzgsbh5rffa

A provenance framework for data-dependent process analysis

Daniel Deutch, Yuval Moskovitch, Val Tannen
2014 Proceedings of the VLDB Endowment  
They can also extend the analysis to incorporate the propagation of annotations from meta-domains of interest, e.g., cost or access privileges.  ...  These constructions address two interacting new challenges: (1) to combine provenance annotations for both information that resides in the database and information about external inputs (e.g., user choices  ...  For the dense structure we have varied both the fan-out and number of levels, and report results fan-out value of 10 and increasing number of levels.  ... 
doi:10.14778/2732279.2732283 fatcat:emm5264rbnhlhbyluh2ndqck2e

Comparative grammar: Concepts and theories (nine Fudan Intensive Summer Teaching lectures) [article]

Martin Haspelmath
2022 Zenodo  
Lecture slides for a three-day FIST (Fudan Intensive Summer Teaching, Shanghai) course, June-July 2022.  ...  Chinese de áγria tis Marías i γáta 'Maria's wild cat' the wild of Maria the cat the burn-t house 毁 c. áγria ke shāohuǐ de fángzi meγáli γáta 'wild and big cat' wild and big cat the man who burnt the house  ...  'pocket', 'Tüte') -meanings described by a picture (of a bag, or of a kind of pepper) or by some other kind of nonverbal stimulus -meanings occurring in parallel texts (e.g. in Bible translations, Wälchli  ... 
doi:10.5281/zenodo.6794792 fatcat:tl42qicaxvbs5hbsami7zwhw24
« Previous Showing results 1 — 15 out of 5,785 results