| 1. |
- Bosch, Nathan, et al.
(författare)
-
Fine-Tuning BERT-based Language Models for Duplicate Trouble Report Retrieval
- 2022
-
Ingår i: Proceedings. - : Institute of Electrical and Electronics Engineers (IEEE). ; , s. 4737-4745
-
Konferensbidrag (refereegranskat)abstract
- In large software-intensive organizations, trouble reports (TRs) are heavily involved in reporting, analyzing, and resolving faults. Due to the scale of modern organizations and products, multiple people independently often identify faults, leading to duplicate TRs. To mitigate the additional manual effort to identify and resolve these duplicate TRs, prior work at Ericsson focused on developing a 2-stage BERT-based retrieval system for identifying similar TRs when provided a new fault observation. This approach, although powerful, struggled to generalize to out-of-domain TRs. In this paper, we evaluate several fine-tuning strategies to integrate domain knowledge further, notably telecommunications knowledge, into the BERT-based TR retrieval models to (i) attain better performance on duplicate TR retrieval/identification and (ii) improve model generalizability to out-of-domain TR data. We find that adding domain-specific data into the fine-tuning models led to improved results on both overall model performance and model generalizability.
|
|
| 2. |
- Shalmashi, Serveh, et al.
(författare)
-
Closed-Form Optimality Characterization of Network-Assisted Device-to-Device Communications
- 2014
-
Ingår i: Proceedings of the IEEE Wireless Communications and Networking Conference (WCNC).
-
Konferensbidrag (refereegranskat)abstract
- This paper considers the mode selection problem for network-assisted device-to-device (D2D) communications with multiple antennas at base station. We study transmission in both dedicated and shared frequency bands. Given the type of resources (i.e., dedicated or shared), the user equipment (UE) decides to transmit in the conventional cellular mode or directly to its corresponding receiver in D2D mode. We formulate this problem under two different objectives. The first problem is to maximize the quality-of-service (QoS) given a transmit power, and the second problem is to minimize the transmit power given a QoS requirement. We derive closed-form results for the optimal decision and show that the two problem formulations behave differently. Taking a geometrical approach, we study the area around the transmitter UE where the receiving UE should be to have D2D mode optimality, and how it is affected by the transmit power, QoS, and the number of base station antennas.
|
|
| 3. |
- Shalmashi, Serveh, et al.
(författare)
-
Cooperative Device-to-Device Communications in the Downlink of Cellular Networks
- 2014
-
Ingår i: Proceedings of the IEEE Wireless Communications and Networking Conference (WCNC). - 9781479930838 ; , s. 2265-2270
-
Konferensbidrag (refereegranskat)abstract
- We propose a cooperative device-to-device (D2D) communications framework in order to combat the problem of congestion in crowded communication environments. The idea is to allow a D2D transmitter to act as an in-band relay for a cellular link and at the same time transmit its own data by employing superposition coding in the downlink. Cooperation between the cellular link and D2D transmitter eases down the requirement on the interference. The main benefit of the proposed method is in increasing the number of connections per unit area with the same spectrum usage. It could also be beneficial to offload over-loaded cells. We formulate our problem to minimize the assigned power for cooperation while making sure the cellular user’s performance does not degrade. Our results show that cooperation possibilities and improvement in overall cell capacity increase with the number of cellular users within the cell as well as the cell size.
|
|
| 4. |
- Shalmashi, Serveh
(författare)
-
Cooperative Spectrum Sharing and Device-to-Device Communications
- 2014
-
Licentiatavhandling (övrigt vetenskapligt/konstnärligt)abstract
- The steep growth in the mobile data traffic has gained a lot of attention in recent years. This growth is mainly the result of emerging applications, multimedia services, and revolutions in the device technology. With current deployments and radio resources, operators will not be able to cope with the growing demands. Consequently, there is a need to either provide new resources or increase the efficiency of what is available. Proposed solutions for accommodating growing data traffic are based on improvements in three dimensions: efficient use of radio resources especially the spectrum, technology advancements, and densifying the current infrastructure. In this thesis, we focus on the spectrum dimension. Providing more spectrum is a long-term process. However, increasing the spectrum usage and efficiency can be put rapidly in practice. We discuss potential solutions in the area of spectrum sharing. Among enabling technologies to facilitate spectrum sharing, we consider the cognitive radio and device-to-device (D2D) communications.In order to gain from sharing the spectrum, systems need to somehow deal with extra sources of interference. In the first part of the thesis, we consider a primary-secondary sharing model in cognitive radio networks. We employ the cooperative communication method in order to facilitate the access of the secondary system to the licensed spectrum of the primary system, and therefore increase the spectrum usage. The cooperation between the two systems is formed provided that it is beneficial for the primary system. In this way, the primary users' quality-of-service can be preserved while at the same time the secondary users can access the spectrum. This cooperative approach prevents both systems from concurrent transmissions. As a consequence, the need for interference control techniques are eliminated. We evaluate different models and transmission schemes and optimize the corresponding parameters to quantify the gain resulting from cooperative spectrum sharing.In the second part of the thesis, we consider spectrum sharing within one system between different types of users. This is done in the context of D2D communications where close proximity users can transmit directly to each other. For this type of communications, either dedicated resources are allocated or resources of the cellular users are reused. We first study the feasibility of cooperation between D2D and cellular users and identify the scenarios where it can be beneficial. Then we take on a challenging problem which guarantees the gain from the D2D communication, namely the mode selection. For this problem, we characterize the decision criteria that determines if D2D communication is gainful. Next, we focus on the problem of interference in D2D communications underlaying cellular networks, where the same spectrum is reused in the spatial domain. In such scenarios, the potential gain is determined by how the interference is managed, which in turn depends on the amount of available information at the base station. The more information is required, the more signaling is needed. In this part of the thesis, we address the trade-off between the signaling overhead and the performance of the system and propose a novel approach for interference control which requires very little information on the D2D users.
|
|
| 5. |
- Shalmashi, Serveh, 1979-
(författare)
-
Device-to-Device Communications for Future Cellular Networks : Challenges, Trade-Offs, and Coexistence
- 2015
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- The steep growth in mobile data traffic has gained a lot of attention in recent years. With current infrastructure deployments and radio resources, operators will not be able to cope with the upcoming demands. Consequently, discussions of the next generation of mobile networks, referred to as the fifth generation (5G), have started in both academia and industry. In addition to more capacity, stringent requirements for improving energy efficiency, decreasing delays, and increasing reliability have been envisioned in 5G. Many solutions have been put forward, one of them being device-to-device (D2D) communications where users in close proximity can transmit directly to one another bypassing the base station (BS).In this thesis, we identify trade-offs and challenges of integrating D2D communications into cellular networks and propose potential solutions. To maximize gains from such integration, resource allocation and interference management are key factors. We start by introducing cooperation between D2D and cellular users in order to minimize any interference between the two user types and identifying the scenarios where this cooperation can be beneficial. It is shown that an increase in the number of cellular users within the coverage area and in the size of the cell is associated with a higher probability of cooperation. With this cooperation, we can potentially increase the number of connected devices, reduce the delay, increase the cell sum rate, and offload an overloaded cell.Next, we consider D2D communications underlaying the uplink of cellular networks. In such a scenario, any potential gain from resource sharing (time, frequency, or space) is determined by how the interference is managed. The quality and performance of the interference management techniques depend on the availability of the channel state information (CSI) and the location of nodes as well as the frequency of updates regarding such information. The more information is required, the more signaling is needed, which results in higher power consumption by the users. We investigate the trade-off between the availability of full CSI, which necessitates instantaneous information, and that of limited CSI, which requires infrequent updates. Our results show that with limited CSI, a good performance (in terms of the sum rate of both user types) can be achieved if a small performance loss is tolerated by cellular users. In addition, we propose a novel approach for interference management which only requires the information on the number of D2D users without any knowledge about their CSI. This blind approach can achieve a small outage probability with very low computational complexity when the number of scheduled D2D users is small.We then study the problem of mode selection, i.e., if a user should transmit in the D2D mode or in the conventional cellular mode. We identify the decision criteria for both overlay and underlay scenarios with two different objectives. We find out that the D2D communication is beneficial in macro cells or at cell boundaries. The area in which D2D mode is optimal varies with the objective of the network, transmit power, required quality-of-service, and the number of BS antennas.In the second part of this thesis, we study the effects of integration and coexistence of underlay D2D communications with another promising technology proposed for 5G, namely massive multiple-input-multiple-output (MIMO). Potential benefits of both technologies are known individually, but the possibility of and performance gains from their coexistence are not adequately addressed. We evaluate the performance of this hybrid network in terms of energy efficiency and the average sum rate. Comprehensive analysis reveals that the performance highly depends on the D2D user density. We conclude that underlay D2D communications can only coexist with massive MIMO systems in the regime of low D2D user density. By introducing a high number of D2D users, gains from the massive MIMO technology degrade rapidly, and therefore in this case, the D2D communications should use the overlay approach rather than the underlay, or the network should only allow a subset of D2D transmissions to be active at a time.
|
|
| 6. |
- Shalmashi, Serveh, et al.
(författare)
-
Energy Efficiency and Sum Rate Tradeoffs for Massive MIMO Systems with Underlaid Device-to-Device Communications
- 2016
-
Ingår i: EURASIP Journal on Wireless Communications and Networking. - : Springer. - 1687-1472 .- 1687-1499.
-
Tidskriftsartikel (refereegranskat)abstract
- In this paper, we investigate the coexistence of two technologies that have been put forward for the fifth generation (5G) of cellular networks, namely, network-assisted device-to-device (D2D) communications and massive MIMO (multiple-input multiple-output). Potential benefits of both technologies are known individually, but the tradeoffs resulting from their coexistence have not been adequately addressed. To this end, we assume that D2D users reuse the downlink resources of cellular networks in an underlay fashion. In addition, multiple antennas at the BS are used in order to obtain precoding gains and simultaneously support multiple cellular users using multiuser or massive MIMO technique. Two metrics are considered, namely the average sum rate (ASR) and energy efficiency (EE). We derive tractable and directly computable expressions and study the tradeoffs between the ASR and EE as functions of the number of BS antennas, the number of cellular users and the density of D2D users within a given coverage area. Our results show that both the ASR and EE behave differently in scenarios with low and high density of D2D users, and that coexistence of underlay D2D communications and massive MIMO is mainly beneficial in low densities of D2D users.
|
|
| 7. |
- Shalmashi, Serveh, et al.
(författare)
-
Energy Efficiency and Sum Rate when Massive MIMO meets Device-to-Device Communication
- 2015
-
Ingår i: Proc. of the IEEE International Conference on Communications (ICC). - : Institute of Electrical and Electronics Engineers (IEEE). ; , s. 627-632
-
Konferensbidrag (refereegranskat)abstract
- This paper considers a scenario of short-range communication, known as device-to-device (D2D) communication, where D2D users reuse the downlink resources of a cellular network to transmit directly to their corresponding receivers. In addition, multiple antennas at the base station (BS) are used in order to simultaneously support multiple cellular users using multiuser or massive MIMO. The network model considers a fixed number of cellular users and that D2D users are distributed according to a homogeneous Poisson point process (PPP). Two metrics are studied, namely, average sum rate (ASR) and energy efficiency (EE). We derive tractable expressions and study the tradeoffs between the ASR and EE as functions of the number of BS antennas and density of D2D users for a given coverage area.
|
|
| 8. |
- Shalmashi, Serveh, et al.
(författare)
-
Interference Constrained Device-to-Device Communications
- 2014
-
Ingår i: 2014 IEEE International Conference on Communications, ICC 2014. - : IEEE Computer Society. - 9781479920037 ; , s. 5245-5250
-
Konferensbidrag (refereegranskat)abstract
- This paper considers a scenario in which multiple device-to-device (D2D) users can reuse the uplink resources of a cellular network to transmit directly to their corresponding receivers. The aggregated interference from the D2D users is limited by applying a threshold on the allowable interference in the base station. The problem is solved under two types of constraints, namely, the peak interference and average interference constraints. In the former, we assume that full channel state information (CSI) is available at the base station, and we optimize the allowable transmit power for the D2D users so that the number of coexisting D2D communications is maximized. We further define a quality-of-service constraint for the D2D users. In practice, however, it is difficult to have complete CSI at the base station as it imposes heavy signaling overhead. Therefore, in the latter scenario, we assume that no knowledge about the location of D2D users and their CSI are available at the base station. This approach does not impose any signaling overhead. Our results show that even with no CSI knowledge, we are able to improve the system performance in terms of throughput by allowing coexisting D2D communications while satisfying the cellular user's constraints.
|
|
| 9. |
- Shalmashi, Serveh, et al.
(författare)
-
Interference management for multiple device-to-device communications underlaying cellular networks
- 2013
-
Ingår i: 2013 IEEE 24th International Symposium on Personal Indoor and Mobile Radio Communications (PIMRC). - : IEEE conference proceedings. - 9781467362351 ; , s. 223-227
-
Konferensbidrag (refereegranskat)abstract
- We study the problem of interference management for device-to-device (D2D) communications where multiple D2D users may coexist with one cellular user. The problem is to optimize the transmit power levels of D2D users to maximize the cell throughput while preserving the signal-to-noise-plus-interference ratio (SINR) performance for the cellular user. This is the so-called multi rate power control problem. We investigate the problem under two assumptions, the availability of the instantaneous or average channel state information (CSI) at the base station. In the first case, D2D transmit power levels adapt to fast fading, whereas in the second case, they only adapt to slow fading. In the latter assumption, the cellular user has a maximum outage probability requirement. With numerical results, we study the trade-off between the signaling overhead, that is frequent CSI feedbacks, and the overall system performance, that is the maximum achievable cell capacity, for D2D communications underlying cellular networks.
|
|
| 10. |
- Shalmashi, Serveh, et al.
(författare)
-
On Secondary User Transmission Schemes in Relay-Assisted Cognitive Radio Networks
- 2013
-
Ingår i: 2013 IEEE 77th Vehicular Technology Conference (VTC Spring). - : IEEE. - 9781467363372 ; , s. 6692665-
-
Konferensbidrag (refereegranskat)abstract
- We consider a cooperative cognitive radio network with one primary user and many secondary users. In each transmission frame, one secondary user is selected to act as a relay for the primary user and also transmits its own data. Three different schemes for the secondary user's transmission is studied, namely, time-splitting, superposition coding, and a combined scheme that takes advantages of both time-splitting and superposition coding schemes. We formulate the relay selection problem for each scheme with the objective of maximizing the secondary user's data rate while satisfying a pre-defined gain threshold for the primary user. This constraint provides an incentive for the primary user when allowing the secondary user access its licensed band. We obtain the optimal solution for this problem in each scheme, on which the relay selection is based on. Finally, we compare the performance of different schemes using Monte-Carlo simulations.
|
|
