Disentangling Latent Emotions of Word Embeddings on Complex Emotional Narratives

Yueyi Jiang; Zhengxuan Wu

REVIEW 5 major objections 6 minor 19 references

Disentangling Latent Emotions of Word Embeddings on Complex Emotional Narratives

T0 review · 5 major / 6 minor · reviewed 2026-08-14 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read A valence-trained linear projection of GloVe vectors creates an emotion space in which Joy plus Trust approximates Love and opposite feelings separate.

desk verdict A plausible but under-controlled sketch: the valence-trained projection may explain the emotion-arithmetic result trivially, so the paper needs null controls before the central claim is credible. read the letter →

arxiv 1908.07817 v1 pith:A7HDM2X2 submitted 2019-08-15 cs.CL cs.IRcs.LG

Zhengxuan Wu , Yueyi Jiang This is my paper · ORCID

classification cs.CLcs.IRcs.LG

keywords wordembeddingsGloVeemotionspacearithmeticvalencepredictionwheelofemotionsEmoLexembeddinginterpretability

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper tries to establish that emotional meaning latent in ordinary word vectors can be brought out by a simple learned linear projection. It trains a small autoregressive model to predict continuous positive-to-negative ratings on emotional narrative transcripts, then uses the model's weight matrix to rotate GloVe vectors into a projected emotion space. In that space, words with positive and negative polarity separate into clusters, the eight basic emotions in the EmoLex lexicon become less entangled than in raw GloVe, and the sum of the vectors for Joy and Trust lands close to the vector for Love while moving away from Remorse. If the claim is right, emotion semantics are partly compositional in distributional word vectors, and a valence-trained projection is enough to expose that composition.

What carries the argument

The carrying object is the 300x300 weight matrix Wv learned by the autoregressive valence-prediction model, used as a linear projection from raw GloVe vectors into the emotion space. The same matrix feeds the paper's dimension-scoring algorithm, which combines Wv with the output-layer weight across dimensions and ranks the 300 dimensions by emotional expressivity, isolating the 34th dimension as the strongest; the top two such dimensions already separate unseen positive and negative words drawn from an affective word list. The claim that emotion arithmetic is preserved rests on cosine similarity measured in this Wv-projected space.

What would settle it

Retrain the projection on valence ratings with the temporal order of each transcript shuffled, or on random ratings, and recompute the Joy + Trust versus Love similarity; if the .36 self-similarity and -.19 opposition to Remorse survive a chance-trained projection, the reported emotion arithmetic is not evidence of an emotion space.

Watch

Extended reading notes

Core claim

The paper's central claim is stated directly: through the proposed emotion space, arithmetic of emotions is preserved in the word vectors. Concretely, when GloVe vectors are projected with the learned matrix, the vector sum Joy + Trust behaves like Love: in raw GloVe the cosine self-similarity of this constructed Love vector is only .10, while after projection it rises to .36, and its similarity to the opposite feeling Remorse drops to -.19. The same comparisons across eight EmoLex emotion categories show that same-valence emotion pairs move closer together in the projection while opposite-valence pairs separate, which the paper reads as evidence that the projected space disentangles emotions better than raw GloVe alone.

Load-bearing premise

The load-bearing premise is that a matrix learned to predict positive-versus-negative valence ratings produces a space where all eight basic emotion categories and their vector sums are meaningfully positioned, rather than merely reflecting one good-bad axis.

Editorial extensions

If this is right

Emotion words can be placed on a continuous affective map directly from embeddings, without requiring hand-built sentiment lexicons for each target domain.
Missing or compound emotion labels can be generated by adding basic emotion vectors; the study constructs Love, Optimism, Submission, and Awe from pairs of EmoLex emotion vectors.
A single valence-trained projection improves separation of same-valence emotion pairs, so polarity-sensitive downstream tasks may benefit from projecting embeddings before classification.
Because only a few embedding dimensions carry most emotional signal, emotion models could be made sparser and more interpretable by attending to those dimensions.
The clustering result on out-of-sample words suggests the projection transfers beyond the narrative training vocabulary.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

The paper does not test whether the projection separates emotions that share valence, such as Joy versus Trust versus Anticipation; a natural extension would measure intra-valence separation to see whether the emotion space is more than a one-dimensional good-bad axis.
If the same arithmetic survives in other distributional embedding families, including context-sensitive embeddings, it would suggest emotion compositionality is a general property of distributional semantics rather than a quirk of GloVe; this is testable with the paper's projection recipe.
A practical extension the authors do not pursue is using the emotion directions for controlled text generation or rewriting, where adding a Joy direction to a sentence embedding would be expected to shift its perceived affect.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Short version: this is a workshop-style exploratory paper with a plausible idea and a load-bearing control problem. The specific claim—that a linear projection trained on valence ratings yields an emotion space where Plutchik-style arithmetic (Joy+Trust≈Love) improves over raw GloVe—is new as far as I know, and the paper is honestly written as a preliminary guide. But the central evaluation never tests the one alternative that would sink it: that the valence-trained Wv merely reweights dimensions by valence, which would trivially make positive words cluster with positive and negative with negative. Without a random-projection baseline, a permutation test, or a map fitted to a non-emotional target, the .10→.36 self-similarity rise and −.19 opposite-feeling separation cannot be attributed to emotion disentanglement. The stress-test note is correct that this is the weak load-bearing joint.

What is good: the SEND setup is real, the autoregressive model is simple and reproducible in principle, and the EmoLex evaluation is a reasonable way to probe emotional geometry. The paper is clear about being preliminary and doesn't oversell the model's predictive performance. The dimension-importance idea, while flawed as stated, is worth thinking about because some GloVe dimensions do carry valence signal.

Soft spots, in proportion: there are no error bars or significance tests on the cosine similarities (Sec. 6.3–6.4), which matters because the reported differences are not huge. Figure 1 selects words using the model's own gradients, so the clean separation is partly in-sample. The 'important dimension' ranking (Sec. 6.1) assumes individual GloVe coordinates are meaningful, but word embeddings have rotational invariance, so coordinate-level importance is not well-defined. There is also a concrete inconsistency between Eq. (3), which treats Wv as a 300×300 matrix, and Sec. 6.3 saying 'element-wise multiplication'—if the actual operation is a per-dimension scaling, the valence-reweighting story is even more direct. The abstract's example says the Joy+Trust sum is similar to Optimism, which seems like a typo for Love, but it should be corrected.

Who is this for: affective computing researchers who want a cheap post-hoc emotion projection. The idea has potential, but the evidence currently does not distinguish it from a valence reweighting. I would send it to peer review only if the authors add proper null controls; without them, the main claim is unsecured. With controls, this could become a solid short paper.

Referee Report

5 major / 6 minor

Summary. The paper proposes a method for probing emotions in GloVe word embeddings. The authors train an autoregressive linear model to predict continuous valence ratings from the Stanford Emotional Narratives Dataset (SEND), then interpret the learned weights to rank GloVe dimensions by emotional importance. They project GloVe vectors through the learned matrix Wv to define an 'emotion space,' visualize word clusters by polarity, and evaluate whether EmoLex emotion categories are 'disentangled' in this space. Their main claim is that the projected space preserves Plutchik-style emotion arithmetic, e.g., Love = Joy + Trust, with similarity between the constructed Love vector and the Love centroid rising from 0.10 in raw GloVe to 0.36 in the projected space, while similarity to the opposite feeling Remorse falls to -0.19.

Significance. If the central claim were established, the paper would offer a simple, interpretable linear projection for emotion semantics in word embeddings, with potential applications in affective NLP. The manuscript is clearly written, uses public datasets (SEND, EmoLex, GloVe, LIWC), reports model performance in terms of CCC, and makes its scoring algorithm explicit (Alg. 1). These are genuine strengths. However, the key empirical claims are not yet secured: the projection is trained solely on valence, no null or control model is tested, and the evaluation lacks error bars or significance tests. The manuscript is best seen as a preliminary study whose headline claim outruns the evidence, but the gap is addressable with additional controls and statistical reporting.

major comments (5)

[Sec. 6.4, Fig. 4] The central claim that 'arithmetic of emotions is preserved in the word vectors' is not supported against the simplest alternative explanation: Wv is trained only to predict valence on SEND (Eqs. (3)-(5)), and Plutchik's categories are strongly valenced. A projection that reweights dimensions by valence importance will tend to increase within-valence similarity and decrease cross-valence similarity, which is exactly the block structure in Fig. 3 and the Love versus Remorse separation in Fig. 4. The report gives no random-projection baseline, no permutation test, and no comparison to a map fitted on a non-emotional target. As written, the reported 0.10-to-0.36 rise and -0.19 opposite-feeling similarity cannot be attributed to emotion arithmetic rather than to valence reweighting. Please add at least one control: a random projection matched in scale, a permutation of Wv's rows, or a projection fitted to a non-emotional regression target, and report whether the effect size changes.
[Sec. 6.3, Eq. (3)] There is an unresolved inconsistency in the definition of the projection. Eq. (3) defines Wv as a 300x300 matrix with the output o_t in R^{300}, while Sec. 6.3 says the projection is computed by 'element-wise multiplication of the weight Wv' with raw word vectors. If the actual operation is a per-dimension scaling, the projected space is even more clearly a single valence reweighting, and the interpretation of Wv as a general emotion geometry is weakened. Please specify exactly whether the projection is a full linear map or a diagonal scaling, and correct the dimension of the bias term in Eq. (3) (bh is given as R^{600x300} but should be compatible with the 300-dimensional output). This point is load-bearing because the 'emotion space' is defined by the projection matrix.
[Secs. 6.3-6.4] The quantitative results are reported as point estimates with no variance or significance testing. The analysis says word pairs were 'randomly selected' and 'average cosine similarity scores' were computed, but there is no standard error, confidence interval, bootstrap, or significance test on the reported differences (e.g., 0.10 to 0.36 or -0.19). Because the number of word pairs and the sampling procedure are not described precisely, the reader cannot tell whether these differences are stable or would be expected from random sampling within EmoLex. Please report the number of pairs, the sampling procedure, and uncertainty measures, and provide a permutation or bootstrap test for the key comparisons.
[Sec. 6.2, Fig. 1] The 2D visualization in Fig. 1 selects the top 19 words 'ranked by their gradients of forward propagation' using the model's own scoring, and the top 2 dimensions are also selected by the same model. This selection procedure introduces a circularity: the words and dimensions are chosen to be separable by the model's weights, so the resulting cluster separation is expected by construction. To support the claim that emotional polarity is discoverable from GloVe dimensions, the evaluation should use an independent or pre-specified set of words (e.g., all LIWC emotion words, or a held-out list) and should report the separation achieved by the same two dimensions on words not used in selection.
[Sec. 6.1, Alg. 1] The claim that 'only a few dimensions of the word vectors contributed to expressing emotions' relies on an ad-hoc score defined as abs(sum_i Wv_i * Wo_i) and on the assumption that individual GloVe dimensions are separately meaningful. This assumption is not justified in the paper and is known to be questionable for dense embeddings such as GloVe, where semantic information is distributed across dimensions. Please provide a validity check for the dimension-scoring method, for example by ablating the top-ranked dimensions from the word vectors and measuring the change in valence prediction, or by comparing the selected dimensions against a random selection in a prediction task. Without such a check, the 'important dimension' finding is not established.

minor comments (6)

[Abstract and Sec. 6.4] There is a typo in the abstract: 'with the the embedded vector' should read 'with the embedded vector.'
[Sec. 4] The text says GloVe were 'pre-trained on wikipages'; please give the exact GloVe version and corpus (e.g., Wikipedia 2014 + Gigaword 5, 300d).
[Fig. 2] The two panels of Fig. 2 are not described in the caption; please add enough caption text to make the heatmap and the score distribution self-explanatory.
[Sec. 6.3] The phrase 'element-wise multiplication of the weight Wv' is ambiguous; if it is meant as a matrix-vector product, the text should say so, and if it is a Hadamard product, the dimension of Wv should be stated accordingly.
[Eq. (3)] The bias vector bh is declared as R^{600x300}; it should be a vector of dimension 300 (or as consistent with the equation's output). Please correct the notation.
[References] Reference [12] is marked 'Invited Revision to Journal' and appears to be unpublished; please update to the published version or give a preprint DOI.

Circularity Check

1 steps flagged · score 3.0 of 10

One non-central selection step is circular; the EmoLex emotion-space evaluation is external but valence-confounded.

fitted input called prediction [Sec. 6.2, Fig. 1]
"We used out-of-sample words from LIWC 2007 [14] from which we selected top 19 words ranked by their gradients of forward propagation for positive and negative polarities, respectively. Figure 1 shows that words with positive meaning are well separated from words with negative meaning in this space."

The positive/negative LIWC words are chosen by ranking each word's gradient in the already-trained valence model, and the 2-d plot uses the same model's top-scoring dimensions (Alg. 1). Thus the separation in Fig. 1 is an artifact of the selection criterion: the selected words are, by construction, extremes of the model's valence signal, so re-plotting them in model-derived coordinates cannot fail to show polarity separation. The paper presents this as evidence that 'words can be clustered by emotional polarity,' but the clustering was induced by the same model used to evaluate it rather than discovered independently.

full rationale

The paper's central projection claim (Secs. 6.3-6.4) is not definitionally circular: Wv is trained on SEND valence ratings (Eqs. 3-5), while the disentanglement and emotion-arithmetic tests use EmoLex/Plutchik categories, a benchmark external to the training labels. The SEND dataset and CCC metric are introduced via self-citations ([12], [19]), but these are ordinary dataset/method citations and are not the load-bearing step that makes the central claim true. The one clear circular step is in Sec. 6.2: the LIWC words displayed in Fig. 1 are selected by the model's own gradients and then projected using the same model's dimensions, so the resulting polarity clustering is forced by the selection. The main emotion-arithmetic result has a genuine correctness risk rather than a circularity: because Wv is a valence-trained projection and EmoLex categories are strongly valenced, the reported Love/Remorse separation and within-positive similarity increases may be valence reweighting, but this is a missing-control/interpretation problem, not an identity with the training target.

Assumptions & free parameters 3 free parameters · 4 assumptions · 1 invented entities

The central claim rests on a fitted valence model, a specific projection matrix, and a psychological theory of emotion composition. The projection matrix is the key loaded input: all results about the emotion space are evaluated using a space defined by that fitted matrix, and no external validation or random baseline is provided. The paper also implicitly assumes GloVe coordinate axes are meaningful for the dimension-importance claim.

free parameters (3)

sigma (autoregressive smoothing weight) = 0.84
Learned on SEND (Sec. 5); it controls how much the prediction depends on the previous rating rather than current input, and therefore shapes the weights Wv that define the emotion space.
Projection matrix Wv = Learned 300x300 matrix
Fit to valence ratings in Sec. 4 and reused in Sec. 6.3-6.4 to define the emotion space; the central disentanglement and arithmetic results depend entirely on this fitted matrix.
Top-2 dimension selection and top-19 word selection = 2 and 19
Hand-chosen cutoffs in Sec. 6.1-6.2 for visualizing clustering; results may depend on these arbitrary cutoffs and the words are selected by the model's gradients.

assumptions (4)

ad hoc to paper GloVe dimensions are individually meaningful, so ranking them by model weights identifies emotion-relevant dimensions.
Used in Algorithm 1 and Section 6.1 to declare the 34th dimension most important; word embeddings are generally invariant under rotation, so individual coordinate axes are not semantic without alignment.
domain assumption The valence-trained projection Wv transfers to all eight discrete Plutchik emotions.
Sections 6.3 and 6.4 use Wv, fit to continuous valence, to define an emotion space evaluated on EmoLex's eight emotion categories; the paper gives no evidence that valence supervision constrains eight-way emotion structure.
domain assumption Plutchik's wheel of emotions, e.g. Joy+Trust=Love, is an appropriate ground truth for embedding arithmetic.
Section 6.4 adopts Plutchik [16] as the taxonomy for generated feeling vectors; this is a psychological theory, not an established property of vector spaces, and may not map onto cosine similarity in embeddings.
domain assumption SEND continuous valence ratings, interpolated per word, are reliable enough to train an interpretable emotion projection.
Section 3 describes the interpolation of 0.5s ratings to words; the model's CCC is only 0.35-0.37 versus human 0.46-0.47, so the learned weights inherit annotation noise.

invented entities (1)

Emotion space (projected vector space defined by Wv)
purpose: Provide a coordinate system where Plutchik emotion arithmetic and valence-polarity clustering are visible.
The space is constructed from the fitted valence model; the only evidence for it is the same EmoLex comparison used to evaluate it, with no external falsifiable prediction.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Disentangling Latent Emotions of Word Embeddings on Complex Emotional Narratives." pith.science (2026). https://pith.science/paper/A7HDM2X2

@misc{pith2026190807817,
  author       = {Pith},
  title        = {Pith review of: Disentangling Latent Emotions of Word Embeddings on Complex Emotional Narratives},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/A7HDM2X2}},
  note         = {Machine review of arXiv:1908.07817}
}

read the original abstract

Word embedding models such as GloVe are widely used in natural language processing (NLP) research to convert words into vectors. Here, we provide a preliminary guide to probe latent emotions in text through GloVe word vectors. First, we trained a neural network model to predict continuous emotion valence ratings by taking linguistic inputs from Stanford Emotional Narratives Dataset (SEND). After interpreting the weights in the model, we found that only a few dimensions of the word vectors contributed to expressing emotions in text, and words were clustered on the basis of their emotional polarities. Furthermore, we performed a linear transformation that projected high dimensional embedded vectors into an emotion space. Based on NRC Emotion Lexicon (EmoLex), we visualized the entanglement of emotions in the lexicon by using both projected and raw GloVe word vectors. We showed that, in the proposed emotion space, we were able to better disentangle emotions than using raw GloVe vectors alone. In addition, we found that the sum vectors of different pairs of emotion words successfully captured expressed human feelings in the EmoLex. For example, the sum of two embedded word vectors expressing Joy and Trust which express Love shared high similarity (similarity score .62) with the embedded vector expressing Optimism. On the contrary, this sum vector was dissimilar (similarity score -.19) with the the embedded vector expressing Remorse. In this paper, we argue that through the proposed emotion space, arithmetic of emotions is preserved in the word vectors. The affective representation uncovered in emotion vector space could shed some light on how to help machines to disentangle emotion expressed in word embeddings.

Figures

Figures reproduced from arXiv: 1908.07817 by the authors.

**Figure 2.** Illustration of calculated scores on each dimension of the 300-dimensional [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗

**Figure 3.** Heatmaps of cosine similarities scores between words with paired emotions. (a) [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗

Figures from the paper (1 more)

**Figure 4.** Figure 4: Heatmaps of cosine similarities scores between words with paired feelings. (a) is [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

19 extracted references · 15 canonical work pages

[1]

In: Joint European conference on machine learning and knowledge discov- ery in databases

Bordes, A., Weston, J., Usunier, N.: Open question answering with weakly supervised em- bedding models. In: Joint European conference on machine learning and knowledge discov- ery in databases. pp. 165–180. Springer (2014)

work page 2014
[2]

In: Proceedings of the IEEE conference on computer vision and pattern recogni- tion

Donahue, J., Anne Hendricks, L., Guadarrama, S., Rohrbach, M., Venugopalan, S., Saenko, K., Darrell, T.: Long-term recurrent convolutional networks for visual recognition and de- scription. In: Proceedings of the IEEE conference on computer vision and pattern recogni- tion. pp. 2625–2634 (2015)

work page 2015
[3]

In: Proceedings of the 2015 ACM on International Conference on Multimodal Interaction

Ebrahimi Kahou, S., Michalski, V ., Konda, K., Memisevic, R., Pal, C.: Recurrent neural net- works for emotion recognition in video. In: Proceedings of the 2015 ACM on International Conference on Multimodal Interaction. pp. 467–474. ACM (2015)

work page 2015
[4]

In: Proceedings of the 14th Conference of the European Chapter of the Association for Computational Linguistics

Faruqui, M., Dyer, C.: Improving vector space word representations using multilingual cor- relation. In: Proceedings of the 14th Conference of the European Chapter of the Association for Computational Linguistics. pp. 462–471 (2014)

work page 2014
[5]

arXiv preprint arXiv:1711.02281 (2017)

Gu, J., Bradbury, J., Xiong, C., Li, V .O., Socher, R.: Non-autoregressive neural machine translation. arXiv preprint arXiv:1711.02281 (2017)

arXiv 2017
[6]

In: Proceedings of the 52nd Annual Meeting of the Association for Computational Linguistics (V olume 1: Long Papers)

Ji, Y ., Eisenstein, J.: Representation learning for text-level discourse parsing. In: Proceedings of the 52nd Annual Meeting of the Association for Computational Linguistics (V olume 1: Long Papers). vol. 1, pp. 13–24 (2014)

work page 2014
[7]

arXiv preprint arXiv:1506.01066 (2015)

Li, J., Chen, X., Hovy, E., Jurafsky, D.: Visualizing and understanding neural models in nlp. arXiv preprint arXiv:1506.01066 (2015)

arXiv 2015
[8]

IEEE Transactions on Affective Computing 8(4), 443–456 (2017)

Li, M., Lu, Q., Long, Y ., Gui, L.: Inferring affective meanings of words from word embed- ding. IEEE Transactions on Affective Computing 8(4), 443–456 (2017)

work page 2017

Show all 19 references

[9]

Biometrics pp

Lin, L.I.K.: A concordance correlation coefﬁcient to evaluate reproducibility. Biometrics pp. 255–268 (1989)

1989
[10]

arXiv preprint arXiv:1301.3781 (2013)

Mikolov, T., Chen, K., Corrado, G., Dean, J.: Efﬁcient estimation of word representations in vector space. arXiv preprint arXiv:1301.3781 (2013)

2013 arXiv
[11]

Proceedings of the National Academy of Sciences 114(37), 9843–9847 (2017) Title Suppressed Due to Excessive Length 9

Morelli, S.A., Ong, D.C., Makati, R., Jackson, M.O., Zaki, J.: Empathy and well-being cor- relate with centrality in different social networks. Proceedings of the National Academy of Sciences 114(37), 9843–9847 (2017) Title Suppressed Due to Excessive Length 9

2017
[12]

Ong, D.C., Wu, Z., Zhi-Xuan, T., Reddan, M., Kahhale, I., Mattek, A., Zaki, J.: Modeling emotion in complex stories: the Stanford Emotional Narratives Dataset (Invited Revision to Journal)
[13]

Cognition 143, 141–162 (2015)

Ong, D.C., Zaki, J., Goodman, N.D.: Affective cognition: Exploring lay theories of emotion. Cognition 143, 141–162 (2015)

2015
[14]

Mahway: Lawrence Erlbaum Associates 71(2001), 2001 (2001)

Pennebaker, J.W., Francis, M.E., Booth, R.J.: Linguistic inquiry and word count: Liwc 2001. Mahway: Lawrence Erlbaum Associates 71(2001), 2001 (2001)

2001
[15]

In: Proceedings of the 2014 conference on empirical methods in natural language processing (EMNLP)

Pennington, J., Socher, R., Manning, C.: Glove: Global vectors for word representation. In: Proceedings of the 2014 conference on empirical methods in natural language processing (EMNLP). pp. 1532–1543 (2014)

2014
[16]

In: Theories of emotion, pp

Plutchik, R.: A general psychoevolutionary theory of emotion. In: Theories of emotion, pp. 3–33. Elsevier (1980)

1980
[17]

Behavioral and Brain Sciences 25(1), 1–20 (2002)

Preston, S.D., De Waal, F.B.: Empathy: Its ultimate and proximate bases. Behavioral and Brain Sciences 25(1), 1–20 (2002)

2002
[18]

In: The Thirtieth International Flairs Conference (2017)

Seyeditabari, A., Zadrozny, W.: Can word embeddings help ﬁnd latent emotions in text? preliminary results. In: The Thirtieth International Flairs Conference (2017)

2017
[19]

IEEE Affective Computing and Intelligent Interaction (ACII) (2019)

Wu, Z., Zhang, X., Zhi-Xuan, T., Zaki, J., Ong, D.C.: Attending to emotional narratives. IEEE Affective Computing and Intelligent Interaction (ACII) (2019)

2019

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.

[1] [1]

In: Joint European conference on machine learning and knowledge discov- ery in databases

Bordes, A., Weston, J., Usunier, N.: Open question answering with weakly supervised em- bedding models. In: Joint European conference on machine learning and knowledge discov- ery in databases. pp. 165–180. Springer (2014)

work page 2014

[2] [2]

In: Proceedings of the IEEE conference on computer vision and pattern recogni- tion

Donahue, J., Anne Hendricks, L., Guadarrama, S., Rohrbach, M., Venugopalan, S., Saenko, K., Darrell, T.: Long-term recurrent convolutional networks for visual recognition and de- scription. In: Proceedings of the IEEE conference on computer vision and pattern recogni- tion. pp. 2625–2634 (2015)

work page 2015

[3] [3]

In: Proceedings of the 2015 ACM on International Conference on Multimodal Interaction

Ebrahimi Kahou, S., Michalski, V ., Konda, K., Memisevic, R., Pal, C.: Recurrent neural net- works for emotion recognition in video. In: Proceedings of the 2015 ACM on International Conference on Multimodal Interaction. pp. 467–474. ACM (2015)

work page 2015

[4] [4]

In: Proceedings of the 14th Conference of the European Chapter of the Association for Computational Linguistics

Faruqui, M., Dyer, C.: Improving vector space word representations using multilingual cor- relation. In: Proceedings of the 14th Conference of the European Chapter of the Association for Computational Linguistics. pp. 462–471 (2014)

work page 2014

[5] [5]

arXiv preprint arXiv:1711.02281 (2017)

Gu, J., Bradbury, J., Xiong, C., Li, V .O., Socher, R.: Non-autoregressive neural machine translation. arXiv preprint arXiv:1711.02281 (2017)

arXiv 2017

[6] [6]

In: Proceedings of the 52nd Annual Meeting of the Association for Computational Linguistics (V olume 1: Long Papers)

Ji, Y ., Eisenstein, J.: Representation learning for text-level discourse parsing. In: Proceedings of the 52nd Annual Meeting of the Association for Computational Linguistics (V olume 1: Long Papers). vol. 1, pp. 13–24 (2014)

work page 2014

[7] [7]

arXiv preprint arXiv:1506.01066 (2015)

Li, J., Chen, X., Hovy, E., Jurafsky, D.: Visualizing and understanding neural models in nlp. arXiv preprint arXiv:1506.01066 (2015)

arXiv 2015

[8] [8]

IEEE Transactions on Affective Computing 8(4), 443–456 (2017)

Li, M., Lu, Q., Long, Y ., Gui, L.: Inferring affective meanings of words from word embed- ding. IEEE Transactions on Affective Computing 8(4), 443–456 (2017)

work page 2017