Pith. sign in

REVIEW 3 major objections 5 minor 4 cited by

The Ambiguous World of Emotion Representation

T0 review · 3 major / 5 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 AMBER describes every emotion representation scheme used in affective computing, from single categorical labels to time-varying distributions over numerical and ordinal labels, in one mathematical language.

desk verdict AMBER gives affective computing a genuinely useful common vocabulary for emotion labels, but the Section 7 universality claim is broader than the formalism in Eq. (3) supports. read the letter →

arxiv 1909.00360 v1 pith:H2PVLXJM submitted 2019-09-01 cs.HC cs.AIcs.LG

Vidhyasaharan Sethu , Emily Mower Provost , Julien Epps , Carlos Busso , Nicholas Cummins , Shrikanth Narayanan This is my paper · ORCID
classification cs.HCcs.AIcs.LG
keywords emotionrepresentationaffectivecomputingambiguitycategoricallabelsnumericalordinalannotatordisagreementAMBERframework
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

AMBER is a proposed common mathematical language for emotion labels. It claims that every way affective computing has represented emotion—categorical labels, numerical dimensions such as arousal and valence, ordinal ratings or rankings, and time-varying versions of all three—can be described by the same two ingredients: a finite set of ordered attribute descriptors and an ambiguity function assigning a non-negative number to each element of each descriptor. The point is to make implicit assumptions explicit, so researchers can compare schemes, see what treating annotator disagreement as noise actually assumes, and design representations that carry ambiguity rather than averaging it away. If the claim holds, affective computing gains a rigorous way to reason about ground truth and about conversions between label types.

What carries the argument

The central objects are attribute descriptors and ambiguity functions. Each attribute descriptor $X^{(n)}$ is an ordered set $\{x^{(n)} \mid \alpha^{(n)} \preceq x^{(n)} \preceq \beta^{(n)}\}$: categorical schemes use two-element sets $\{O, I\}$ per emotion category, numerical schemes use intervals such as $[-1,1]$ for valence and arousal, and ordinal schemes use finite ordered sets without any distance notion. The ambiguity function $\xi_{n,t}: X^{(n)} \to [0,\infty)$ assigns a non-negative number to every element of each descriptor at time $t$, encoding how likely that element is; equation (4) shows that single-valued representations are exactly the special case where the function is 1 on one element and 0 elsewhere. This pair of ingredients does the work of unifying the three label spaces and of turning previously hidden assumptions into explicit mathematical restrictions.

What would settle it

Find a published emotion representation whose label space cannot be written as a finite set of ordered attribute descriptors with non-negative per-element ambiguity functions—for example, one whose meaning depends on the joint distribution or correlation between dimensions, such as a valence–arousal trace where the correlation itself carries information—and show that it cannot be reconstructed from the per-descriptor ambiguity functions alone. Such a case would refute the claim that every scheme fits AMBER without adding new structure.

Watch

Extended reading notes

Core claim

The central claim is that the AMBER framework can describe every emotion representation scheme employed in affective computing to date. An emotion at time $t$ is represented as a set of functions $\xi_{n,t}: X^{(n)} \to [0,\infty)$ over attribute descriptors, where each descriptor $X^{(n)}$ is an ordered set with lowest element $\alpha^{(n)}$ and highest element $\beta^{(n)}$. Single-valued categorical, numerical, and ordinal labels are special cases in which the ambiguity function equals 1 for the chosen element and 0 for all others; blended emotions, emotion profiles, distributional labels, and multi-annotator traces are less constrained instances of the same structure. The consequence is that implicit choices—mutual exclusivity of categories, Euclidean distance in the arousal–valence plane, the assumption that annotators transcribe identically, or the assumption that labels do not vary within an interval—can be stated explicitly and compared across schemes.

Load-bearing premise

The framework assumes that every emotion representation can be faithfully captured by a finite number of ordered attribute descriptors, with all ambiguity expressed as a non-negative number attached to each element of each descriptor; if emotions require relations between descriptors, context-dependent meaning, or a structure that an ordered set cannot express, the universality claim fails.

Editorial extensions

If this is right

  • Researchers can express any existing emotion representation in a shared notation, making categorical, numerical, and ordinal schemes directly comparable.
  • Ambiguity-aware training becomes a well-defined task: the loss can be a divergence between the predicted ambiguity function and the target ambiguity function, such as Kullback–Leibler divergence when both are probability distributions.
  • Conversions between schemes—for example, from Likert-type annotations to continuous $[-1,1]$ predictions, or from numerical labels to categorical clusters—can be analyzed for the distance metrics and orthogonality assumptions they silently introduce.
  • Treating the averaged annotation as ground truth is revealed as a special assumption of certainty rather than a neutral default, so datasets and models that ignore inter-annotator disagreement become open to explicit scrutiny.
  • Time-varying ambiguity functions provide a natural language for modeling annotator reaction lag, drift, and systematic disagreement over time.

Reading between the lines

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

  • If the universality claim holds, one could build a representation calculus in which schemes are compared by their assumption load, and a researcher choosing a label scheme could see exactly which mathematical restrictions they are buying into.
  • A testable extension would be to define a formal measure of the representational capacity of a scheme—for instance, the number of independent ambiguity functions or the richness of the ordered sets—and use it to predict when one annotation scheme can substitute for another.
  • The framework suggests that future affective benchmarks might report full ambiguity functions instead of averaged labels, making inter-annotator disagreement a first-class evaluation target rather than noise to be removed.
  • One could extend AMBER to compare ambiguity functions across modalities, treating each modality as a separate attribute descriptor and using the framework to reason about multimodal emotion perception.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. The paper introduces the AMBiguous Emotion Representation (AMBER) framework, a mathematical language intended to describe categorical, numerical, and ordinal emotion representation schemes used in affective computing, including time-varying and ambiguity-aware variants. The framework represents an emotion at time t as a set of elements chosen from a finite number of ordered attribute descriptor sets, and extends this to per-descriptor ambiguity functions. The authors reinterpret existing schemes, including categorical labels, emotion profiles, blended emotions, dimensional labels, annotation distributions, and ordinal/rank labels, through AMBER, and argue in Section 7 that every emotion representation scheme employed in affective computing to date can be expressed in the framework. The paper is a conceptual contribution with no empirical evaluation.

Significance. If the universality claim could be made precise, AMBER would give affective computing researchers a common vocabulary for exposing and comparing the implicit assumptions of emotion labels, addressing a real need in the field. The paper's strengths are the clarity of the per-descriptor formalism, the worked re-descriptions in Boxes 1 through 4 and Section 5, and the explicit separation of ambiguity (perception and expression) from uncertainty (model imprecision). The central limitation is that the framework's core definition in Eq. (3) only captures per-descriptor ambiguity, and the paper does not formally account for joint structure across attribute descriptors; this gap is load-bearing because the universality claim rests on it.

major comments (3)
  1. [Section 4.2, Eq. (3)] The definition of Ψ(t) as a set of N independent functions ξ_{n,t} on each descriptor X^(n) cannot represent joint ambiguity across descriptors. For example, the two representations 'either happy or sad with probability 0.5 each' and 'happy and sad each independently present with probability 0.5' produce identical ambiguity functions under Eq. (3): ξ_happy(I) = ξ_sad(I) = 0.5 and ξ_happy(O) = ξ_sad(O) = 0.5 in both cases, even though they encode different distributions over label combinations. Since the framework introduces special constraints such as Eqs. (9)-(11) only for the blended-emotions example, it lacks a general mechanism for expressing cross-descriptor constraints. This makes the Section 7 universality claim either false (if only Eq. (3) is intended) or unfalsifiable (if arbitrary constraints are allowed). Please specify how joint and relational structure is represented in the general framework, or qualify the universality claim accordingly.
  2. [Section 7] The statement 'every emotion representation scheme employed in affective computing till date can be described within the proposed AMBER framework' is asserted after only a few illustrative examples. No enumeration of representation families is provided, and the term 'described' is not given a formal meaning. In particular, it is not clear what class of constraints such as Eqs. (9)-(11) is permitted, and whether the framework is closed under their composition. The paper should either provide a systematic account of the space of emotion representations and a proof sketch of universality, or replace the universal claim with a weaker statement about the examples presented.
  3. [Section 5.3] The proposed ambiguity-aware treatment of relative ordinal labels is not an instance of Eq. (3). The text in Section 5.3 and Figure 6B suggests defining an ambiguity function over the outcomes of each pairwise comparison, but Eq. (3) defines functions on a single descriptor X^(n), not on pairs of time points or comparisons. Similarly, the discussion of temporal dynamics in Section 5.2 mentions constraints between Ψ(t) and Ψ(t-ε) only as an open question. The framework needs a formal account of how such relational and temporal structures are expressed within AMBER.
minor comments (5)
  1. [Section 4.2] The text says ξ maps to '[0, ∞)' but then describes it as associating each element with a 'positive real number'; since Eq. (4) assigns the value 0, the term should be 'non-negative real number'.
  2. [Section 4.1] The requirement that every X^(n) be an ordered set is applied to categorical labels, for which the two elements {O, I} have no natural order. The paper imposes an arbitrary order without discussing whether this order carries any semantic meaning.
  3. [Section 1 and Section 4.2] The distinction between ambiguity and uncertainty introduced in Section 1 is never formally reflected in the framework; the same ambiguity function appears to serve for both annotator disagreement and model prediction uncertainty. The paper should clarify whether ξ can represent one, the other, or both, and how the distinction is operationalized.
  4. [Section 5.1] The text contains a minor typo: 'an more general' should read 'a more general' when referring to the emotion profile representation.
  5. [Box 1] The Iverson bracket notation in Eq. (10) is defined immediately after its first use; moving the definition before the equation would improve readability. Additionally, the phrase '[condi t i on]' has unusual spacing in the preprint.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: AMBER is a proposed descriptive framework rather than a fitted predictive derivation, and the examples drawn from the authors' prior work are illustrative, not load-bearing evidence.

full rationale

The paper does not fit parameters and then predict a closely related quantity, and it invokes no uniqueness theorem or external formal result that is assumed from a self-citation. The central mathematical content is definitional: Eq. (3) defines an ambiguity-aware emotion representation as a set of per-descriptor functions, with Eq. (2) and Eq. (4) as the unambiguous special case. This is a proposed modelling language, not a derivation that converts an input into an output. Section 7's universality claim ('every emotion representation scheme employed in affective computing till date can be described within the proposed AMBER framework') is an assertion about scope; it is not entailed by the definitions and may be over-broad (Eq. (3) does not by itself formalise joint ambiguity across descriptors), but an unsupported or over-broad scope claim is a correctness risk, not circularity. The self-citations (emotion profiles [49], distribution-based labels [19,30], relative ordinal labels [54]) are used only to illustrate how existing schemes fit the framework; no derived quantity is subsequently predicted from those examples. The paper also identifies open questions, such as the unspecified form of temporal constraints, which further indicates that it is not presenting a forced derivation. The paper is self-contained as a proposal, so the appropriate finding is no significant circularity.

Assumptions & free parameters 0 free parameters · 3 assumptions · 2 invented entities

The framework's central claim rests on the expressive power of ordered attribute descriptors and ambiguity functions. These are introduced as definitions, not derived, and their adequacy for all emotion label spaces is assumed rather than shown. No free parameters are fitted to data; the constructs are conceptual.

assumptions (3)
  • domain assumption A finite number of attribute descriptors, each an ordered set with least and greatest elements, suffices to describe any emotion label space.
    Section 4.1, Equation (1) posits N ordered sets X(n) spanning categorical, numerical, and ordinal labels. The paper provides examples but no proof that all possible label spaces have this structure.
  • ad hoc to paper Emotional ambiguity can be quantified by assigning a non-negative number to every possible label element via an ambiguity function.
    Section 4.2, Equation (3) defines ξn,t: X(n) → [0,∞). This is the paper's modeling choice; it assumes ambiguity is representable as per-attribute non-negative scores, which excludes certain relational or holistic ambiguity notions.
  • domain assumption Existing emotion representation schemes are faithfully preserved when re-expressed in AMBER.
    Section 5 maps categorical, numerical, and ordinal schemes into AMBER; the mapping assumes the ordering and distance properties of the original schemes are correctly captured by the attribute descriptor sets and functions.
invented entities (2)
  • Attribute descriptors (ordered sets X(n))
    purpose: Provide the common substrate for categorical, numerical, and ordinal emotion label spaces.
    This is the paper's proposed building block; no external empirical evidence is offered that real emotion label spaces decompose this way.
  • Ambiguity function ξn,t
    purpose: Quantify the degree to which each possible label value is supported at time t.
    Defined in Section 4.2; it is a flexible construct that reduces to point estimates when it is 1 for one element and 0 otherwise. It has no falsifiable handle outside the framework.

how reviews work

0 comments
Cite this review

Pith. "Pith review of The Ambiguous World of Emotion Representation." pith.science (2026). https://pith.science/paper/H2PVLXJM

@misc{pith2026190900360,
  author       = {Pith},
  title        = {Pith review of: The Ambiguous World of Emotion Representation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/H2PVLXJM}},
  note         = {Machine review of arXiv:1909.00360}
}
read the original abstract

Artificial intelligence and machine learning systems have demonstrated huge improvements and human-level parity in a range of activities, including speech recognition, face recognition and speaker verification. However, these diverse tasks share a key commonality that is not true in affective computing: the ground truth information that is inferred can be unambiguously represented. This observation provides some hints as to why affective computing, despite having attracted the attention of researchers for years, may not still be considered a mature field of research. A key reason for this is the lack of a common mathematical framework to describe all the relevant elements of emotion representations. This paper proposes the AMBiguous Emotion Representation (AMBER) framework to address this deficiency. AMBER is a unified framework that explicitly describes categorical, numerical and ordinal representations of emotions, including time varying representations. In addition to explaining the core elements of AMBER, the paper also discusses how some of the commonly employed emotion representation schemes can be viewed through the AMBER framework, and concludes with a discussion of how the proposed framework can be used to reason about current and future affective computing systems.

Figures

Figures reproduced from arXiv: 1909.00360 by the authors.

Figure 1
Figure 1. An illustration of the Brunswik functional lens model depicting the expression and perception of emotions [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Categorical labels of sentences from the MSP-Podcast corpus on the arousal-valence space. The categorical [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 3
Figure 3. A depiction of the differences between ambiguous and non-ambiguous categorical emotion representa [PITH_FULL_IMAGE:figures/full_fig_p010_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: (A) A time-varying numerical emotion representations within the proposed AMBER framework based on [PITH_FULL_IMAGE:figures/full_fig_p011_4.png]
Figure 5
Figure 5. Figure 5: Deriving relative labels from time-continuous traces using the qualitative agreement analysis [54]. (A) The [PITH_FULL_IMAGE:figures/full_fig_p013_5.png]
Figure 6
Figure 6. Figure 6: (A) Absolute ordinal representation may be made ambiguity aware by defining a distribution over the finite [PITH_FULL_IMAGE:figures/full_fig_p014_6.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 4 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Indeterminacy in Affective Computing: Considering Meaning and Context in Data Collection Practices

    cs.AI 2025-02 conditional novelty 6.0 of 10

    Affective labels inherit four qualities of indeterminacy from human interpretation, and the paper argues that data collection must track the context of those interpretations to make emotion prediction reliable.

  2. Rethinking Emotion Annotations in the Era of Large Language Models

    cs.CL 2024-12 conditional novelty 6.0 of 10

    Human evaluators preferred GPT-4's zero-shot emotion labels over original human labels in 62% of disagreement samples, and GPT-4 pre-filtering and post-filtering can reduce annotation workload and improve training efficiency.

  3. Meta-PerSER: Few-Shot Listener Personalized Speech Emotion Recognition via Meta-learning

    eess.AS 2025-05 conditional novelty 5.0 of 10

    Meta-PerSER uses MAML-style meta-training with combined-set training, derivative annealing, and per-layer learning rates to personalize speech emotion recognition to unseen annotators from 32 labeled examples, outperf...

  4. Emotions as Ambiguity-aware Ordinal Representations

    cs.LG 2025-08 conditional novelty 4.0 of 10

    Ordinal representations based on the rate of change of ambiguous emotion annotations improve prediction of directional changes in continuous emotion traces, especially for unbounded labels like engagement.

Reference graph

Works this paper leans on

79 extracted references · 79 canonical work pages · cited by 4 Pith papers

  1. [1]

    K., and Calvo, R

    AlZoubi, O., D’Mello, S. K., and Calvo, R. A. (2012). Detecting naturalistic expressions of nonbasic affect using physiological signals. IEEE Transactions on Affective Computing, 3(3):298–310

    work page 2012

  2. [2]

    Atcheson, M., Sethu, V ., and Epps, J. (2017). Gaussian process regression for continuous emotion recognition with global temporal invariance. In IJCAI 2017 Workshop on Artificial Intelligence in Affective Computing, pages 34–44

    work page 2017

  3. [3]

    Atcheson, M., Sethu, V ., and Epps, J. (2018). Demonstrating and modelling systematic time-varying annotator disagreement in continuous emotion annotation. Proc. Interspeech 2018, pages 3668–3672

    work page 2018

  4. [4]

    Batliner, A., Steidl, S., and Nöth, E. (2008). Releasing a thoroughly annotated and processed spontaneous emotional database: the FAU Aibo emotion corpus. In Second International Workshop on Emotion: Corpora for 15 THE AMBIGUOUS WORLD OF EMOTION REPRESENTATION (P REPRINT ) Research on Emotion and Affect, International conference on Language Resources and ...

    work page 2008

  5. [5]

    and Gruber, A

    Bickmore, T . and Gruber, A. (2010). Relational Agents in Clinical Psychiatry. Harvard Review of Psychiatry, 18(2):119–130

    work page 2010

  6. [6]

    and Schulman, D

    Bickmore, T . and Schulman, D. (2007). Practical Approaches to Comforting Users with Relational Agents. InCHI ’07 Extended Abstracts on Human Factors in Computing Systems, pages 2291–2296, San Jose, CA, USA. ACM

    work page 2007

  7. [7]

    Burges, C., Shaked, T ., Renshaw, E., Lazier, A., Deeds, M., Hamilton, N., and Hullender, G. (2005). Learning to rank using gradient descent. In International conference on Machine learning (ICML 2005), pages 89–96, Bonn, Germany

    work page 2005

  8. [8]

    Busso, C., Parthasarathy, S., Burmania, A., AbdelWahab, M., Sadoughi, N., and Mower Provost, E. (2017). MSP- IMPROV: An acted corpus of dyadic interactions to study emotion perception. IEEE Transactions on Affective Computing, 8(1):67–80

    work page 2017

Show all 79 references
  1. [9]

    Calvo, R. A. and D’Mello, S. (2010). Affect Detection: An Interdisciplinary Review of Models, Methods, and Their Applications. IEEE Transactions on Affective Computing, 1(1):18–37

    2010

  2. [10]

    Calvo, R. A. and D’Mello, S. (2012). Frontiers of Affect-Aware Learning Technologies.IEEE Intelligent Systems, 27(6):86–89

    2012

  3. [11]

    Chao, L., Tao, J., Yang, M., Li, Y., and Wen, Z. (2015). Long short term memory recurrent neural network based multimodal dimensional emotion recognition. In Proceedings of the 5th International Workshop on Audio/Visual Emotion Challenge, pages 65–72. ACM

    2015

  4. [12]

    and Kuncheva, L

    Christy, T . and Kuncheva, L. I. (2014). Technological advancements in affective gaming: A historical survey. GSTF Journal on Computing, 3(4):34–41

    2014

  5. [13]

    and Cornelius, R

    Cowie, R. and Cornelius, R. (2003). Describing the emotional states that are expressed in speech. Speech Communication, 40(1-2):5–32

    2003

  6. [14]

    Cowie, R., Douglas-Cowie, E., Savvidou, S., McMahon, E., Sawey, M., and Schröder, M. (2000). ’FEELTRACE’: An instrument for recording perceived emotion in real time. In ISCA Tutorial and Research Workshop (ITRW) on Speech and Emotion, pages 19–24, Newcastle, Northern Ireland, UK. ISCA

    2000

  7. [15]

    Cowie, R., Douglas-Cowie, E., Tsapatsoulis, N., Votsis, G., Kollias, S., Fellenz, W ., and Taylor, J. G. (2001). Emotion recognition in human-computer interaction. IEEE Signal processing magazine, 18(1):32–80

    2001

  8. [16]

    Cowie, R., Sawey, M., Doherty, C., Jaimovich, J., Fyans, C., and Stapleton, P . (2013). Gtrace: General trace program compatible with EmotionML. In Affective Computing and Intelligent Interaction (ACII 2013), pages 709–710, Geneva, Switzerland

    2013

  9. [17]

    Cummins, N., Scherer, S., Krajewski, J., Schnieder, S., Epps, J., and Quatieri, T . F . (2015). A review of depression and suicide risk assessment using speech analysis. Speech Communication, 71:10–49

    2015

  10. [18]

    Dang, T ., Sethu, V ., and Ambikairajah, E. (2018). Dynamic multi-rater gaussian mixture regression incorporating temporal dependencies of emotion uncertainty using kalman filters. In 2018 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pages ...

    2018

  11. [19]

    Dang, T ., Sethu, V ., Epps, J., and Ambikairajah, E. (2017). An investigation of emotion prediction uncertainty using gaussian mixture regression. In INTERSPEECH, pages 1248–1252

    2017

  12. [20]

    Devillers, L., Vidrascu, L., and Lamel, L. (2005). Challenges in real-life emotion annotation and machine learning based detection. Neural Networks, 18(4):407–422

    2005

  13. [21]

    D’Mello, S. K. and Graesser, A. C. (2014). Feeling, thinking, and computing with affect-aware learning. In Calvo, R., D’Mello, S., Gratch, J., and Kappas, A., editors, The Oxford Handbook of Affective Computing, chapter 31, pages 419–434. Oxford University Press

    2014

  14. [22]

    M., and Vogel, C

    Esposito, A., Esposito, A. M., and Vogel, C. (2015). Needs and challenges in human computer interaction for processing social emotional information. Pattern Recognition Letters, 66:41–51

    2015

  15. [23]

    M., Lech, M., and Cavedon, L

    Fayek, H. M., Lech, M., and Cavedon, L. (2016). Modeling subjectiveness in emotion recognition with deep neural networks: Ensembles vs soft labels. In Neural Networks (IJCNN), 2016 International Joint Conference on, pages 566–570. IEEE

    2016

  16. [24]

    Frijda, N. H. (1986). The emotions. Cambridge University Press

    1986

  17. [25]

    Girard, J. M. (2014). CARMA: Software for continuous affect rating and media annotation. Journal of Open Research Software, 2(1):1–6. 16 THE AMBIGUOUS WORLD OF EMOTION REPRESENTATION (P REPRINT )

    2014

  18. [26]

    and Kroschel, K

    Grimm, M. and Kroschel, K. (2005). Evaluation of natural emotions using self assessment manikins. In IEEE Workshop on Automatic Speech Recognition and Understanding, 2005., pages 381–385. IEEE

    2005

  19. [27]

    Grimm, M., Kroschel, K., Mower, E., and Narayanan, S. (2007). Primitives-based evaluation and estimation of emotions in speech. Speech Communication, 49(10-11):787–800

    2007

  20. [28]

    and Schuller, B

    Gunes, H. and Schuller, B. (2013). Categorical and dimensional affect analysis in continuous input: Current trends and future directions. Image and Vision Computing, 31(2):120–136

    2013

  21. [29]

    Gunes, H., Schuller, B., Pantic, M., and Cowie, R. (2011). Emotion representation, analysis and synthesis in continuous space: A survey. In Automatic Face & Gesture Recognition and Workshops (FG 2011), 2011 IEEE International Conference on, pages 827–834. IEEE

    2011

  22. [30]

    Han, J., Zhang, Z., Schmitt, M., Pantic, M., and Schuller, B. (2017). From hard to soft: Towards more human-like emotion recognition by modelling the perception uncertainty. In Proceedings of the 25th ACM international conference on Multimedia, pages 890–897. ACM

    2017

  23. [31]

    M., Lajoie, S

    Harley, J. M., Lajoie, S. P ., Frasson, C., and Hall, N. C. (2017). Developing Emotion-Aware, Advanced Learning Technologies: A Taxonomy of Approaches and Features. International Journal of Artificial Intelligence in Education, 27(2):268–297

    2017

  24. [32]

    and Sebe, N

    Jaimes, A. and Sebe, N. (2007). Multimodal human-computer interaction: A survey. Computer Vision and Image Understanding, 108(1):116–134

    2007

  25. [33]

    Joachims, T . (2006). Training linear SVMs in linear time. InACM SIGKDD international conference on Knowledge discovery and data mining, pages 217–226, Philadelphia, USA

    2006

  26. [34]

    Z., Li, J., and Luo, J

    Joshi, D., Datta, R., Fedorovskaya, E., Luong, Q.-T ., Wang, J. Z., Li, J., and Luo, J. (2011). Aesthetics and emotions in images. IEEE Signal Processing Magazine, 28(5):94–115

    2011

  27. [35]

    Khorram, S., Jaiswal, M., Gideon, J., McInnis, M., and Provost, E.-M. (2018). The priori emotion dataset: Linking mood to emotion detected in-the-wild. Interspeech 2018, pages 1903–1907

    2018

  28. [36]

    Khorram, S., McInnis, M., and Provost, E.-M. (2019). Jointly aligning and predicting continuous emotion annotations. IEEE Transactions on Affective Computing, page To appear

    2019

  29. [37]

    and Kim, J

    Kim, Y. and Kim, J. (2018). Human-like emotion recognition: Multi-label learning from noisy labeled audio- visual expressive speech. In 2018 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pages 5104–5108. IEEE

    2018

  30. [38]

    C., Baucom, B

    Lee, C.-C., Black, M., Katsamanis, A., Lammert, A. C., Baucom, B. R., Christensen, A., Georgiou, P . G., and Narayanan, S. S. (2010). Quantification of prosodic entrainment in affective spontaneous spoken interactions of married couples. In Eleventh Annual Conference of the Int...

    2010

  31. [39]

    and Busso, C

    Lotfian, R. and Busso, C. (2017). Formulating emotion perception as a probabilistic model with application to categorical emotion classification. In International Conference on Affective Computing and Intelligent Interaction (ACII 2017), pages 415–420, San Antonio, TX, USA

    2017

  32. [40]

    and Busso, C

    Lotfian, R. and Busso, C. (2018). Predicting categorical emotions by jointly learning primary and secondary emotions through multitask learning. In Interspeech 2018, pages 951–955, Hyderabad, India

    2018

  33. [41]

    and Busso, C

    Lotfian, R. and Busso, C. (2019). Building naturalistic emotionally balanced speech corpus by retrieving emotional speech from existing podcast recordings. IEEE Transactions on Affective Computing, To appear

    2019

  34. [42]

    and Busso, C

    Mariooryad, S. and Busso, C. (2013). Analysis and compensation of the reaction lag of evaluators in continuous emotional annotations. In Affective Computing and Intelligent Interaction (ACII 2013), pages 85–90, Geneva, Switzerland

    2013

  35. [43]

    and Busso, C

    Mariooryad, S. and Busso, C. (2015). Correcting time-continuous emotional labels by modeling the reaction lag of evaluators. IEEE Transactions on Affective Computing, 6(2):97–108. Special Issue Best of ACII

    2015

  36. [44]

    Martinez, H., Yannakakis, G., and Hallam, J. (2014). Don’t classify ratings of affect; rank them!IEEE transactions on affective computing, (1):1–1

    2014

  37. [45]

    G., and Mower Provost, E

    Matton, K., McInnis, M. G., and Mower Provost, E. (2019). Into the wild: Transitioning from recognizing mood in clinical interactions to personal conversations for individuals with bipolar disorder. In Interspeech

    2019

  38. [46]

    McKeown, G., Valstar, M., Cowie, R., Pantic, M., and Schröder, M. (2012). The SEMAINE database: Anno- tated multimodal records of emotionally colored conversations between a person and a limited agent. IEEE Transactions on Affective Computing, 3(1):5–17

    2012

  39. [47]

    and Russell, J

    Mehrabian, A. and Russell, J. A. (1974). An approach to environmental psychology. the MIT Press. 17 THE AMBIGUOUS WORLD OF EMOTION REPRESENTATION (P REPRINT )

    1974

  40. [48]

    S., Duvivier, L

    Messinger, D. S., Duvivier, L. L., Warren, Z., Mahoor, M., Baker, J., Warlaumont, A. S., and Ruvolo, P . (2014). Affective Computing, Emotional Development, and Autism. In Calvo, R., D’Mello, S., Gratch, J., and Kappas, A., editors, The Oxford Handbook of Affective Computing, ...

    2014

  41. [49]

    J., and Narayanan, S

    Mower, E., Mataric, M. J., and Narayanan, S. (2011). A framework for automatic human emotion classification using emotion profiles. IEEE Transactions on Audio, Speech, and Language Processing, 19(5):1057–1070

    2011

  42. [50]

    J., and Narayanan, S

    Mower, E., Matari´ c, M. J., and Narayanan, S. S. (2009). Evaluating evaluators: A case study in understanding the benefits and pitfalls of multi-evaluator modeling. In Tenth Annual Conference of the International Speech Communication Association

    2009

  43. [51]

    Nicolle, J., Rapp, V ., Bailly, K., Prevost, L., and Chetouani, M. (2012). Robust continuous prediction of human emotions using multiscale dynamic cues. In International conference on Multimodal interaction (ICMI 2012), pages 501–508, Santa Monica, CA, USA

    2012

  44. [52]

    L., and Collins, A

    Ortony, A., Clore, G. L., and Collins, A. (1990). The cognitive structure of emotions. Cambridge university press

    1990

  45. [53]

    Pampouchidou, A., Simos, P ., Marias, K., Meriaudeau, F ., Yang, F ., Pediaditis, M., and Tsiknakis, M. (2017). Automatic assessment of depression based on visual cues: A systematic review. IEEE Transactions on Affective Computing

    2017

  46. [54]

    Parthasarathy, S., Cowie, R., and Busso, C. (2016). Using agreement on direction of change to build rank-based emotion classifiers. IEEE/ACM Transactions on Audio, Speech, and Language Processing, 24(11):2108–2121

    2016

  47. [55]

    Picard, R. W . (2000). Affective computing. MIT press

    2000

  48. [56]

    Picard, R. W . and Healey, J. (1997). Affective wearables. Personal Technologies, 1(4):231–240

    1997

  49. [57]

    Plutchik, R. (1980). Emotion: A Psychoevolutionary Synthesis. Harper and Row

    1980

  50. [58]

    Politou, E., Alepis, E., and Patsakis, C. (2017). A survey on mobile affective computing. Computer Science Review, 25:79–100

    2017

  51. [59]

    Riva, G., Calvo, R., and Lisetti, C. (2014). Cyberpsychology and Affective Computing. In Calvo, R., D’Mello, S., Gratch, J., and Kappas, A., editors, The Oxford Handbook of Affective Computing, chapter 41, pages 547–558. Oxford University Press

    2014

  52. [60]

    Russell, J. A. (1980). A circumplex model of affect. Journal of personality and social psychology, 39(6):1161

    1980

  53. [61]

    Russell, J. A. (1993). Forced-choice response format in the study of facial expression. Motivation and Emotion, 17(1):41–51

    1993

  54. [62]

    Scherer, K. (1984). On the nature and function of emotion: A component process approach , pages 293–317. Lawrence Erlbaum Associates, Inc., New Jersey

    1984

  55. [63]

    Scherer, K. (2003). Vocal communication of emotion: A review of research paradigms. Speech Communication, 40(1-2):227–256

    2003

  56. [64]

    Schlosberg, H. (1941). A scale for the judgment of facial expressions. Journal of experimental psychology , 29(6):497

    1941

  57. [65]

    Schlosberg, H. (1954). Three dimensions of emotion. Psychological review, 61(2):81

    1954

  58. [66]

    Schröder, M., Baggia, P ., Burkhardt, F ., Pelachaud, C., Peter, C., and Zovato, E. (2015). Emotion markup language. The Oxford Handbook of Affective Computing, page 395

    2015

  59. [67]

    Schröder, M., Cowie, R., Douglas-Cowie, E., Westerdijk, M., and Gielen, S. (2001). Acoustic correlates of emotion dimensions in view of speech synthesis. In Seventh European Conference on Speech Communication and Technology

    2001

  60. [68]

    Schröder, M., Devillers, L., Karpouzis, K., Martin, J.-C., Pelachaud, C., Peter, C., Pirker, H., Schuller, B., Tao, J., and Wilson, I. (2007). What should a generic emotion markup language be able to represent? In International Conference on Affective Computing and Intelligent...

    2007

  61. [69]

    Schuller, B., Valster, M., Eyben, F ., Cowie, R., and Pantic, M. (2012). Avec 2012: the continuous audio/visual emotion challenge. In Proceedings of the 14th ACM international conference on Multimodal interaction, pages 449–456. ACM

    2012

  62. [70]

    and Robinson, P

    Sobol-Shikler, T . and Robinson, P . (2010). Classification of complex information: Inference of co-occurring affective states from their expressions in speech. IEEE Transactions on Pattern Analysis and Machine Intelligence, 32(7):1284–1297

    2010

  63. [71]

    and Devillers, L

    Vidrascu, L. and Devillers, L. (2005). Real-life emotion representation and detection in call centers data. In International Conference on Affective Computing and Intelligent Interaction, pages 739–746. Springer. 18 THE AMBIGUOUS WORLD OF EMOTION REPRESENTATION (P REPRINT )

    2005

  64. [72]

    F ., Cristani, M., Fuhrmann, F ., Gilmartin, E., Hammal, Z., Heylen, D., Kaiser, R., Koutsombogera, M., Potamianos, A., Renals, S., Riccardi, G., and Salah, A

    Vinciarelli, A., Esposito, A., André, E., Bonin, F ., Chetouani, M., Cohn, J. F ., Cristani, M., Fuhrmann, F ., Gilmartin, E., Hammal, Z., Heylen, D., Kaiser, R., Koutsombogera, M., Potamianos, A., Renals, S., Riccardi, G., and Salah, A. A. (2015). Open Challenges in Modelling...

    2015

  65. [73]

    Wang, J.-C., Yang, Y.-H., Wang, H.-M., and Jeng, S.-K. (2015). Modeling the affective content of music with a gaussian mixture model. IEEE Transactions on Affective Computing, 6(1):56–68

    2015

  66. [74]

    and Tellegen, A

    Watson, D. and Tellegen, A. (1985). Toward a consensual structure of mood. Psychological bulletin, 98(2):219

    1985

  67. [75]

    Yannakakis, G., Cowie, R., and Busso, C. (2017). The ordinal nature of emotions. In International Conference on Affective Computing and Intelligent Interaction (ACII 2017), pages 248–255, San Antonio, TX, USA

    2017

  68. [76]

    Yannakakis, G., Cowie, R., and Busso, C. (2019). The ordinal nature of emotions: An emerging approach. IEEE Transactions on Affective Computing, To appear

    2019

  69. [77]

    Yannakakis, G. N. and Paiva, A. (2014). Emotion in Games. In Calvo, R., D’Mello, S., Gratch, J., and Kappas, A., editors, The Oxford Handbook of Affective Computing, chapter 34, pages 458–471. Oxford University Press

    2014

  70. [78]

    Zhang, B., Essl, G., and Mower Provost, E. (2017). Predicting the distribution of emotion perception: capturing inter-rater variability. In Proceedings of the 19th ACM International Conference on Multimodal Interaction, pages 51–59. ACM

    2017

  71. [79]

    Zhou, Y., Xue, H., and Geng, X. (2015). Emotion distribution recognition from facial expressions. In Proceedings of the 23rd ACM international conference on Multimedia, pages 1247–1250. ACM. 19

    2015

Pith tools

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