Exploration and generalization in deep learning with SwitchPath activations
This work provides a comprehensive theoretical and empirical analysis of SwitchPath, a stochastic activation function that improves learning dynamics by probabilistically toggling between a neuron standard activation and its negation. We develop theoretical foundations and demonstrate its impact in...
Saved in:
| Published in: | Machine learning Vol. 114; no. 9; p. 200 |
|---|---|
| Main Authors: | , , , , , , |
| Format: | Journal Article |
| Language: | English |
| Published: |
Dordrecht
Springer Nature B.V
01.09.2025
|
| Subjects: | |
| ISSN: | 0885-6125, 1573-0565, 1573-0565 |
| Online Access: | Get full text |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Abstract | This work provides a comprehensive theoretical and empirical analysis of SwitchPath, a stochastic activation function that improves learning dynamics by probabilistically toggling between a neuron standard activation and its negation. We develop theoretical foundations and demonstrate its impact in multiple scenarios. By maintaining gradient flow and injecting controlled stochasticity, the method improves generalization, uncertainty estimation, and training efficiency. Experiments in classification show consistent gains over ReLU and Leaky ReLU across CNNs and Vision Transformers, with reduced overfitting and better test accuracy. In generative modeling, a novel two-phase training scheme significantly mitigates mode collapse and accelerates convergence. Our theoretical analysis reveals that SwitchPath introduces a form of multiplicative noise that acts as a structural regularizer. Additional empirical investigations show improved information propagation and reduced model complexity. These results establish this activation mechanism as a simple yet effective way to enhance exploration, regularization, and reliability in modern neural networks. |
|---|---|
| AbstractList | This work provides a comprehensive theoretical and empirical analysis of SwitchPath, a stochastic activation function that improves learning dynamics by probabilistically toggling between a neuron standard activation and its negation. We develop theoretical foundations and demonstrate its impact in multiple scenarios. By maintaining gradient flow and injecting controlled stochasticity, the method improves generalization, uncertainty estimation, and training efficiency. Experiments in classification show consistent gains over ReLU and Leaky ReLU across CNNs and Vision Transformers, with reduced overfitting and better test accuracy. In generative modeling, a novel two-phase training scheme significantly mitigates mode collapse and accelerates convergence. Our theoretical analysis reveals that SwitchPath introduces a form of multiplicative noise that acts as a structural regularizer. Additional empirical investigations show improved information propagation and reduced model complexity. These results establish this activation mechanism as a simple yet effective way to enhance exploration, regularization, and reliability in modern neural networks. |
| ArticleNumber | 200 |
| Author | Morandin, Francesco Parton, Maurizio Di Cecco, Antonio Fantozzi, Marco Metta, Carlo Papini, Andrea Galfrè, Silvia Giulia |
| Author_xml | – sequence: 1 givenname: Antonio surname: Di Cecco fullname: Di Cecco, Antonio – sequence: 2 givenname: Andrea surname: Papini fullname: Papini, Andrea – sequence: 3 givenname: Carlo surname: Metta fullname: Metta, Carlo – sequence: 4 givenname: Marco surname: Fantozzi fullname: Fantozzi, Marco – sequence: 5 givenname: Silvia Giulia surname: Galfrè fullname: Galfrè, Silvia Giulia – sequence: 6 givenname: Francesco surname: Morandin fullname: Morandin, Francesco – sequence: 7 givenname: Maurizio surname: Parton fullname: Parton, Maurizio |
| BackLink | https://research.chalmers.se/publication/547849$$DView record from Swedish Publication Index (Chalmers tekniska högskola) |
| BookMark | eNo9kUtLxTAQhYMoeH38AVcF19VJ0yTtUsQXCgrqeojJxFupaU161euvN1pxMzMMhzNn-HbYZhgCMXbA4YgD6OPEoW3rEipZgmpqKNcbbMGlFiVIJTfZAppGlopXcpvtpPQCAJVq1IJdn32O_RDN1A2hMMEVzxQomr77mlddKBzRWPRkYujCc_HRTcviPle7vDN5NHbq3n-1aY9tedMn2v_ru-zx_Ozh9LK8ub24Oj25Ka2AeiprQW1rrCLTGidrK4lU40nLHMl6r6GqW-dsJZTwqiElK8-5kEaT97JxVuyy-9k3fdC4esIxdq8mrnEwHUZKOaldol2a_pViwkQIwjnBpUAveYO1B4UmX0SwlRbWSq1dm10PZ9cxDm8rShO-DKsY8iMochbdgoIfVTWrbBxSiuT_r3PAHxQ4o8CMAn9R4Fp8A4CYf5Q |
| Cites_doi | 10.1007/978-3-030-03243-2_860-1 10.1007/978-3-031-58547-0_5 10.1109/TSP49548.2020.9163446 10.1038/nature16961 10.1109/ICCV.2015.123 10.1023/A:1022621522141 10.1007/978-3-642-16111-7_23 10.1007/978-3-031-78977-9_18 10.1109/ICMLA55696.2022.00099 10.1038/nature14236 10.1109/IJCNN.2019.8852266 10.1016/j.neuron.2016.06.033 10.1134/S0006297924603915 10.1109/WACV57701.2024.00279 10.1038/nn.4243 10.1016/j.isci.2023.107963 10.1007/s10994-006-6226-1 10.1007/978-3-319-46493-0_38 10.1093/oso/9780195121049.001.0001 10.1023/A:1010933404324 10.1038/sdata.2018.161 10.1109/ISCC53001.2021.9631485 10.1145/279943.279989 10.3233/FAIA200119 10.1523/JNEUROSCI.0410-21.2021 10.1109/ICCV.2015.172 10.1145/3422622 10.1007/978-3-319-46493-0_39 10.1038/s41467-022-30305-8 10.1126/science.1211095 10.1016/j.cviu.2018.10.009 10.1145/2939672.2939778 10.1109/CVPR42600.2020.00813 10.1038/nrn2258 |
| ContentType | Journal Article |
| Copyright | The Author(s), under exclusive licence to Springer Science+Business Media LLC, part of Springer Nature 2025. |
| Copyright_xml | – notice: The Author(s), under exclusive licence to Springer Science+Business Media LLC, part of Springer Nature 2025. |
| DBID | AAYXX CITATION 7SC 8FD JQ2 L7M L~C L~D ADTPV AOWAS F1S |
| DOI | 10.1007/s10994-025-06840-y |
| DatabaseName | CrossRef Computer and Information Systems Abstracts Technology Research Database ProQuest Computer Science Collection Advanced Technologies Database with Aerospace Computer and Information Systems Abstracts Academic Computer and Information Systems Abstracts Professional SwePub SwePub Articles SWEPUB Chalmers tekniska högskola |
| DatabaseTitle | CrossRef Computer and Information Systems Abstracts Technology Research Database Computer and Information Systems Abstracts – Academic Advanced Technologies Database with Aerospace ProQuest Computer Science Collection Computer and Information Systems Abstracts Professional |
| DatabaseTitleList | Computer and Information Systems Abstracts |
| DeliveryMethod | fulltext_linktorsrc |
| Discipline | Computer Science |
| EISSN | 1573-0565 |
| ExternalDocumentID | oai_research_chalmers_se_03dd3153_f518_4f06_a8fe_0c273cc577d9 10_1007_s10994_025_06840_y |
| GroupedDBID | -Y2 -~C -~X .4S .86 .DC .VR 06D 0R~ 0VY 199 1N0 1SB 2.D 203 28- 29M 2J2 2JN 2JY 2KG 2KM 2LR 2P1 2VQ 2~H 30V 4.4 406 408 409 40D 40E 5GY 5QI 5VS 67Z 6NX 6TJ 78A 88I 8AO 8FE 8FG 8TC 8UJ 95- 95. 95~ 96X AAAVM AABHQ AACDK AAEWM AAHNG AAIAL AAJBT AAJKR AANZL AAOBN AAPKM AARHV AARTL AASML AATNV AATVU AAUYE AAWCG AAYIU AAYQN AAYTO AAYXX ABAKF ABBBX ABBRH ABBXA ABDBE ABDZT ABECU ABFSG ABFTV ABHLI ABHQN ABIVO ABJNI ABJOX ABKCH ABKTR ABMNI ABMQK ABNWP ABQBU ABQSL ABRTQ ABSXP ABTEG ABTHY ABTKH ABTMW ABULA ABUWG ABWNU ABXPI ACAOD ACBXY ACDTI ACGFS ACGOD ACHSB ACHXU ACKNC ACMDZ ACMLO ACNCT ACOKC ACOMO ACPIV ACSTC ACZOJ ADHHG ADHIR ADHKG ADIMF ADKFA ADKNI ADKPE ADMLS ADRFC ADTPH ADURQ ADYFF ADZKW AEBTG AEFIE AEFQL AEGAL AEGNC AEJHL AEJRE AEKMD AEMSY AENEX AEOHA AEPYU AETLH AEVLU AEXYK AEZWR AFBBN AFDZB AFEXP AFFHD AFGCZ AFHIU AFKRA AFLOW AFOHR AFQWF AFWTZ AFZKB AGAYW AGDGC AGJBK AGMZJ AGQEE AGQMX AGQPQ AGRTI AGWIL AGWZB AGYKE AHAVH AHBYD AHKAY AHPBZ AHSBF AHWEU AHYZX AIAKS AIGIU AIIXL AILAN AITGF AIXLP AJBLW AJRNO AJZVZ ALMA_UNASSIGNED_HOLDINGS ALWAN AMKLP AMVHM AMXSW AMYLF AMYQR AOCGG ARAPS ARCSS ARMRJ ASPBG ATHPR AVWKF AXYYD AYFIA AYJHY AZFZN AZQEC B-. BA0 BBWZM BDATZ BENPR BGLVJ BGNMA BPHCQ BSONS CAG CCPQU CITATION COF CS3 CSCUP DDRTE DL5 DNIVK DPUIP DU5 DWQXO EBLON EBS EIOEI EJD ESBYG F5P FEDTE FERAY FFXSO FIGPU FINBP FNLPD FRRFC FSGXE FWDCC GGCAI GGRSB GJIRD GNUQQ GNWQR GQ7 GQ8 GXS H13 HCIFZ HF~ HG5 HG6 HMJXF HQYDN HRMNR HVGLF HZ~ I-F I09 IHE IJ- IKXTQ ITG ITH ITM IWAJR IXC IZIGR IZQ I~X I~Y I~Z J-C J0Z JBSCW JCJTX JZLTJ K6V K7- KDC KOV KOW LAK LLZTM M2P M4Y MA- MVM N2Q N9A NB0 NDZJH NPVJJ NQJWS NU0 O9- O93 O9G O9I O9J OAM OVD P19 P2P P62 P9O PF- PHGZM PHGZT PQGLB PQQKQ PROAC PT4 Q2X QF4 QM1 QN7 QO4 QOK QOS R4E R89 R9I RHV RNI RNS ROL RPX RSV RZC RZE S16 S1Z S26 S27 S28 S3B SAP SCJ SCLPG SCO SDH SHX SISQX SJYHP SNE SNPRN SNX SOHCF SOJ SPISZ SRMVM SSLCW STPWE SZN T13 T16 TAE TEORI TN5 TSG TSK TSV TUC TUS U2A UG4 UOJIU UTJUX UZXMN VC2 VFIZW W23 W48 WH7 WIP WK8 YLTOR Z45 Z8Z ZMTXR 7SC 8FD AESKC JQ2 L7M L~C L~D ADTPV AOWAS F1S |
| ID | FETCH-LOGICAL-c304t-43e99ac6ea9ad54c5ee68fe75026cff70249ddc2363f68e652f1135a7eff58dc3 |
| IEDL.DBID | RSV |
| ISICitedReferencesCount | 1 |
| ISICitedReferencesURI | http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=001544702500002&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D |
| ISSN | 0885-6125 1573-0565 |
| IngestDate | Wed Nov 05 04:21:09 EST 2025 Wed Nov 05 09:04:37 EST 2025 Sat Nov 29 07:36:42 EST 2025 |
| IsPeerReviewed | true |
| IsScholarly | true |
| Issue | 9 |
| Language | English |
| LinkModel | DirectLink |
| MergedId | FETCHMERGED-LOGICAL-c304t-43e99ac6ea9ad54c5ee68fe75026cff70249ddc2363f68e652f1135a7eff58dc3 |
| Notes | ObjectType-Article-1 SourceType-Scholarly Journals-1 ObjectType-Feature-2 content type line 14 |
| PQID | 3236790609 |
| PQPubID | 54194 |
| ParticipantIDs | swepub_primary_oai_research_chalmers_se_03dd3153_f518_4f06_a8fe_0c273cc577d9 proquest_journals_3236790609 crossref_primary_10_1007_s10994_025_06840_y |
| PublicationCentury | 2000 |
| PublicationDate | 2025-09-01 |
| PublicationDateYYYYMMDD | 2025-09-01 |
| PublicationDate_xml | – month: 09 year: 2025 text: 2025-09-01 day: 01 |
| PublicationDecade | 2020 |
| PublicationPlace | Dordrecht |
| PublicationPlace_xml | – name: Dordrecht |
| PublicationTitle | Machine learning |
| PublicationYear | 2025 |
| Publisher | Springer Nature B.V |
| Publisher_xml | – name: Springer Nature B.V |
| References | 6840_CR19 6840_CR11 6840_CR55 6840_CR57 6840_CR14 6840_CR58 6840_CR15 AA Faisal (6840_CR18) 2008; 9 6840_CR59 6840_CR16 PL Bartlett (6840_CR5) 2002; 3 6840_CR50 6840_CR51 6840_CR53 RS Sutton (6840_CR64) 2018 6840_CR10 6840_CR54 D Silver (6840_CR60) 2016; 529 C Gao (6840_CR21) 2016; 59 MI Ajibola (6840_CR1) 2021; 41 S Dutta (6840_CR17) 2022; 13 P Smyth (6840_CR62) 1994; 15 N Srivastava (6840_CR63) 2014; 15 V Mnih (6840_CR52) 2015; 518 6840_CR44 G Amato (6840_CR100) 2024; 3762 6840_CR45 6840_CR46 6840_CR48 6840_CR49 JR Quinlan (6840_CR56) 1993; 10 6840_CR41 6840_CR42 6840_CR43 ER Kandel (6840_CR36) 2000 R Tremblay (6840_CR66) 2016; 91 B Yang (6840_CR72) 2023; 26 A Borji (6840_CR8) 2019; 179 6840_CR33 6840_CR34 6840_CR3 6840_CR35 6840_CR2 6840_CR37 6840_CR4 6840_CR7 6840_CR39 6840_CR6 SA Kauffman (6840_CR38) 2000 6840_CR70 6840_CR71 6840_CR73 6840_CR31 VF Kitchigina (6840_CR40) 2025; 90 6840_CR32 P Tschandl (6840_CR67) 2018; 5 P Geurts (6840_CR23) 2006; 63 L Breiman (6840_CR9) 2001; 45 I Goodfellow (6840_CR25) 2020; 63 6840_CR22 6840_CR24 C Metta (6840_CR47) 2023; 16 6840_CR69 6840_CR26 TP Vogels (6840_CR68) 2011; 334 6840_CR27 6840_CR28 6840_CR29 S Deneve (6840_CR12) 2016; 19 6840_CR61 6840_CR20 6840_CR65 G Hennequin (6840_CR30) 2017; 94 A Destexhe (6840_CR13) 2012; 106 |
| References_xml | – volume: 94 start-page: 816 issue: 4 year: 2017 ident: 6840_CR30 publication-title: Neuron – ident: 6840_CR51 doi: 10.1007/978-3-030-03243-2_860-1 – ident: 6840_CR57 – volume: 59 issue: 3 year: 2016 ident: 6840_CR21 publication-title: Science China Information Sciences – ident: 6840_CR11 – ident: 6840_CR14 doi: 10.1007/978-3-031-58547-0_5 – volume: 106 start-page: 5 issue: 1–2 year: 2012 ident: 6840_CR13 publication-title: Journal of Physiology-Paris – ident: 6840_CR4 – ident: 6840_CR39 doi: 10.1109/TSP49548.2020.9163446 – volume: 529 start-page: 484 issue: 7587 year: 2016 ident: 6840_CR60 publication-title: Nature doi: 10.1038/nature16961 – volume: 3 start-page: 463 year: 2002 ident: 6840_CR5 publication-title: Journal of Machine Learning Research – ident: 6840_CR28 doi: 10.1109/ICCV.2015.123 – ident: 6840_CR44 – volume: 15 start-page: 169 issue: 2 year: 1994 ident: 6840_CR62 publication-title: Machine Learning doi: 10.1023/A:1022621522141 – ident: 6840_CR65 doi: 10.1007/978-3-642-16111-7_23 – ident: 6840_CR15 doi: 10.1007/978-3-031-78977-9_18 – ident: 6840_CR35 – ident: 6840_CR55 doi: 10.1109/ICMLA55696.2022.00099 – ident: 6840_CR10 – volume: 518 start-page: 529 issue: 7540 year: 2015 ident: 6840_CR52 publication-title: Nature doi: 10.1038/nature14236 – ident: 6840_CR73 – ident: 6840_CR54 doi: 10.1109/IJCNN.2019.8852266 – volume: 91 start-page: 260 issue: 2 year: 2016 ident: 6840_CR66 publication-title: Neuron doi: 10.1016/j.neuron.2016.06.033 – volume: 3762 start-page: 494 year: 2024 ident: 6840_CR100 publication-title: CEUR Workshop Proceedings – ident: 6840_CR31 – volume: 90 start-page: 61 issue: 1 year: 2025 ident: 6840_CR40 publication-title: Biochemistry (Moscow) doi: 10.1134/S0006297924603915 – ident: 6840_CR48 doi: 10.1109/WACV57701.2024.00279 – volume: 16 start-page: 243 issue: 3 year: 2023 ident: 6840_CR47 publication-title: International Journal of Data Science and Analytics – volume: 19 start-page: 375 issue: 3 year: 2016 ident: 6840_CR12 publication-title: Nature Neuroscience doi: 10.1038/nn.4243 – ident: 6840_CR45 – ident: 6840_CR3 – ident: 6840_CR20 – volume: 10 start-page: 271 issue: 3 year: 1993 ident: 6840_CR56 publication-title: Machine Learning – ident: 6840_CR41 – ident: 6840_CR7 – ident: 6840_CR24 – volume: 26 start-page: 107963 issue: 10 year: 2023 ident: 6840_CR72 publication-title: Iscience doi: 10.1016/j.isci.2023.107963 – volume: 63 start-page: 3 issue: 1 year: 2006 ident: 6840_CR23 publication-title: Machine Learning doi: 10.1007/s10994-006-6226-1 – ident: 6840_CR29 doi: 10.1007/978-3-319-46493-0_38 – volume-title: Investigations year: 2000 ident: 6840_CR38 doi: 10.1093/oso/9780195121049.001.0001 – volume: 45 start-page: 5 issue: 1 year: 2001 ident: 6840_CR9 publication-title: Machine Learning doi: 10.1023/A:1010933404324 – volume: 5 start-page: 180161 year: 2018 ident: 6840_CR67 publication-title: Scientific Data doi: 10.1038/sdata.2018.161 – ident: 6840_CR49 doi: 10.1109/ISCC53001.2021.9631485 – ident: 6840_CR59 – volume-title: Reinforcement learning: An introduction year: 2018 ident: 6840_CR64 – ident: 6840_CR70 – ident: 6840_CR46 doi: 10.1145/279943.279989 – ident: 6840_CR32 – ident: 6840_CR53 doi: 10.3233/FAIA200119 – ident: 6840_CR69 – volume: 41 start-page: 8181 issue: 39 year: 2021 ident: 6840_CR1 publication-title: Journal of Neuroscience doi: 10.1523/JNEUROSCI.0410-21.2021 – ident: 6840_CR61 – ident: 6840_CR2 – ident: 6840_CR27 – ident: 6840_CR6 – volume-title: Principles of neural science year: 2000 ident: 6840_CR36 – ident: 6840_CR42 doi: 10.1109/ICCV.2015.172 – volume: 63 start-page: 139 issue: 11 year: 2020 ident: 6840_CR25 publication-title: ACM doi: 10.1145/3422622 – ident: 6840_CR34 doi: 10.1007/978-3-319-46493-0_39 – ident: 6840_CR71 – volume: 13 start-page: 1899 issue: 1 year: 2022 ident: 6840_CR17 publication-title: Nature Communications doi: 10.1038/s41467-022-30305-8 – ident: 6840_CR33 – ident: 6840_CR19 – ident: 6840_CR16 – ident: 6840_CR50 – volume: 334 start-page: 1569 issue: 6062 year: 2011 ident: 6840_CR68 publication-title: Science doi: 10.1126/science.1211095 – ident: 6840_CR43 – volume: 179 start-page: 41 year: 2019 ident: 6840_CR8 publication-title: Computer Vision and Image Understanding doi: 10.1016/j.cviu.2018.10.009 – ident: 6840_CR58 doi: 10.1145/2939672.2939778 – ident: 6840_CR37 doi: 10.1109/CVPR42600.2020.00813 – volume: 9 start-page: 292 issue: 4 year: 2008 ident: 6840_CR18 publication-title: Nature Reviews Neuroscience doi: 10.1038/nrn2258 – ident: 6840_CR26 – volume: 15 start-page: 1929 issue: 56 year: 2014 ident: 6840_CR63 publication-title: Journal of Machine Learning Research – ident: 6840_CR22 |
| SSID | ssj0002686 |
| Score | 2.4745138 |
| Snippet | This work provides a comprehensive theoretical and empirical analysis of SwitchPath, a stochastic activation function that improves learning dynamics by... |
| SourceID | swepub proquest crossref |
| SourceType | Open Access Repository Aggregation Database Index Database |
| StartPage | 200 |
| SubjectTerms | Adaptability Algorithms Back propagation Decision making Decision trees Deep learning Empirical analysis Generative networks Gradient flow Machine learning Neural network algorithms Neural networks Neurons Regularization |
| Title | Exploration and generalization in deep learning with SwitchPath activations |
| URI | https://www.proquest.com/docview/3236790609 https://research.chalmers.se/publication/547849 |
| Volume | 114 |
| WOSCitedRecordID | wos001544702500002&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D |
| hasFullText | 1 |
| inHoldings | 1 |
| isFullTextHit | |
| isPrint | |
| journalDatabaseRights | – providerCode: PRVAVX databaseName: SpringerLINK Contemporary 1997-Present customDbUrl: eissn: 1573-0565 dateEnd: 99991231 omitProxy: false ssIdentifier: ssj0002686 issn: 1573-0565 databaseCode: RSV dateStart: 19970101 isFulltext: true titleUrlDefault: https://link.springer.com/search?facet-content-type=%22Journal%22 providerName: Springer Nature |
| link | http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV1LSwUhFJaIFm16R7cXLtqV4PgaXUYUQRFBD9qJo8dumymaW3D_ferM7QG1aDNL0e94jgfH7_sQOuAamGI1kLSBGyKYV0Q7J4kIlTMhiLop_Ir7y_rqSj88mOs5dPTnH_xMcivytUxmgxRByTQV3Er1ZK2b-8-yy1SxdUxZI0k-tgeGzO9D_DyFvrWW3-VCyxFztvy_ya2gpaGVxMd97FfRHLRraHlm04CHrF1HF_07uxIC7NqAH3up6YGBiZ9aHABe8GAg8Yjz3Sy-SV8_vk4NIs7ch_7mtttAd2entyfnZPBQIJ5TMSGCgzHOK3DGBSm8BFA6QuoTmPIx1lkxMATPuOJRaVCSxari0tUQo9TB80003z63sIVwrZvYUCFCE7kADQYUa6iOnDYJe0lH6HCGqX3ppTLslyhyxskmnGzByU5HaHcGux3SprM868kZqqgZocs-FJ9DZR3sQQBpbP24uMt0tgNLeQg8FW4bZaWtiFRZl5ZoqU-tmfeyroPZ_tfcdtAiKyHNT8l20fzk9Q320IJ_nzx1r_tl230AFRXRHw |
| linkProvider | Springer Nature |
| openUrl | ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&rft.genre=article&rft.atitle=Exploration+and+generalization+in+deep+learning+with+SwitchPath+activations&rft.jtitle=Machine+learning&rft.au=Di+Cecco%2C+Antonio&rft.au=Papini%2C+Andrea&rft.au=Metta%2C+Carlo&rft.au=Fantozzi%2C+Marco&rft.date=2025-09-01&rft.pub=Springer+Nature+B.V&rft.issn=0885-6125&rft.eissn=1573-0565&rft.volume=114&rft.issue=9&rft.spage=200&rft_id=info:doi/10.1007%2Fs10994-025-06840-y&rft.externalDBID=HAS_PDF_LINK |
| thumbnail_l | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=0885-6125&client=summon |
| thumbnail_m | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=0885-6125&client=summon |
| thumbnail_s | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=0885-6125&client=summon |