IP Library › Granted Patent US 12,407,444

Granted Patent B1

US 12,407,444 · App. 19/021,669 · Granted Sep 2, 2025

Enhanced reliable communications including streaming codes for partial bursts and guardspaces and synergized compression

Inventor: Michael H. Rudow (Cleveland, OH)

H04L1/0063 H04L1/0041 H04L1/0059

View Patent ↗

Loading inventors, assignments & file history…

Monitor This Case

Get email alerts when status or documents change.

Order Certified Copies

Most orders are placed with the USPTO same day — all within 24 business hours.

Order via The Patent Place →

Pre-filled with this patent's details

Quick Facts

Patent No.: US 12,407,444
App. No.: 19/021,669
Granted: Sep 2, 2025
Kind: B1

External Links

USPTO Patent Center↗ Full Patent PDF↗

Abstract

Enhanced reliable communications systems, methods, computer program products, and integrated circuits include streaming codes for partial burst and guardspaces and synergized compression. An FEC encoder can generate two or more types of parity symbols to address partial burst and guardspace losses. An FEC encoder can utilize information from a data compressor to make frame splitting, parity symbol generation, and/or packetizing decisions. A data compressor can use information from an FEC encoder to make data compression decisions. Aspects of an FEC encoder and/or data compressor can be trained using machine learning, including reinforcement learning. Multimodal operation provides flexibility for dynamically reacting to changing communication conditions.

Claims (334)

1. A frame splitting encoder system comprising:

a frame splitter configured to split each of a number of data frames i into a plurality of components including at least a first component Γ[i] and a second component γ[i];

a parity symbol generator configured to allocate parity symbols for the components to ensure that (a) if the fraction of packets lost during time slot i is at most and the data of all prior frames is available, then the parity symbols sent during a time slot suffice to recover the lost data during that time slot, and (b) if time slot i is part of a partial burst starting in time slot j (i.e., for each time slot z∈{j, . . . , j+b j −1}, l z or fewer fraction of the packets sent during the time slot are lost where i∈{j, . . . , j+b j −1}) followed by a partial guard space (i.e., for each time slot z∈{j+b j , . . . , j+b j −1+τ}, l z (G) or fewer fraction of the packets sent during the time slot are lost) and all the data of frames before the start of the partial burst is available, then the data for frame i is recovered within τ time slots; and

a packetizer configured to packetize the components and the parity symbols.

2. The system of claim 1 , wherein parity symbols are allocated such that all lost data for the first component of the frames is to be recovered by (τ−1) time slots after the start of the partial burst and the second component of each frame of the partial burst is recovered τ time slots later.

3. The system of claim 2 , wherein the second component of the frames in a partial guard space after the partial burst are recovered by (τ−1) time slots after the start of the partial burst excluding future frames τ or more time slots after the start of the partial burst.

4. The system of claim 1 , wherein parity symbols are allocated such that there are two or more types of parity symbols where one type of parity symbol for a frame is independent of the symbols of the second component of the same frame.

5. The system of claim 4 , wherein the parity symbols allocated for a data frame i include at least a first set of parity symbols P[i] and a second set of parity symbols G[i] based on the plurality of components, wherein the second set of parity symbols is configured to ensure that (a) some parity symbols of a data frame can be used to recover the first component of the same data frame, (b) some parity symbols of a data frame can be used to recover the second component of the same data frame, and (c) some parity symbols of a data frame cannot be used to recover the second component of the data frame.

6. The system of claim 5 , wherein the number of parity symbols allocated to P[i] and G[i] is based on the parameter l i (G) .

7. The system of claim 6 , wherein the parameter l i (G) is set to approximate an upper bound on the fraction of packets that may be lost during time slot i if there is no partial burst encompassing time slot i for which loss recovery is needed.

8. The system of claim 6 , wherein l i (G) is set so that the fraction of losses during time slot i is no more than l i (G) with some probability.

9. The system of claim 5 , wherein allocating the parity symbols comprises a two-stage parity allocation of (a) pre-allocating p i+τ during time slot i for robustness to partial bursts then (b) increasing the size of p i during time slot i for robustness to loss in the partial guard space; the size of G[i] can then be set during time slot i.

10. The system of claim 5 , wherein the frame splitting encoder system is multimodal.

11. The system of claim 5 , wherein the number of parity symbols to be sent with the data of data frame i is set with the intention of (a) loss recovery during time slot i if there is no partial burst (and losses defined as partial guard space) and (b) loss recovery during time slot i of the second component of frame (i−τ) (i.e., γ[i−τ]) which reflects if the partial burst encompasses frame (i−τ).

12. The system of claim 11 , wherein:

if the partial burst includes time slot i, then some of the parity symbols are used to recover a subset of Γ[i] and/or some of the parity symbols are used to recover Γ[i−τ+1:i−1]; and

if the partial burst starts after time slot (i−τ) and ends before time slot i, then the parity symbols are used to recover lost symbols of Γ[i−τ+1:i−1] depending on the time slot in which the burst starts.

13. The system of claim 11 , wherein the number of parity symbols p i +g i is set so that when l i (G) fraction of the parity symbols are lost and a partial burst encompasses time slot (i−τ), the number of received parity symbols p i R +g i R is approximately equal to (a) the number of missing symbols Γ[i] plus (b) the number of missing symbols of γ[i−τ] less the number of received symbols of G[i−τ], where this subtraction is bounded below by 0.

14. The system of claim 13 , wherein g; is set so that the at least g i R =(1−l i (G) )g i received symbols of G[i] suffice to recover the lost symbols of γ[i], leaving p i R symbols of P[i] to recover the lost symbols of Γ[i] and γ[i−τ].

15. The system of claim 13 , wherein p i is set approximately as follows:

p i (1− l i (G) )=γ i l i (G) +max(0, l i−τ υ i−τ −(1− l i−τ ) g i−τ ),

leading to:

⌈

⁢

(

)

max

⁡

(

⁢

(

)

⁢

)

(

)

⌉

and the ceiling is taken to ensure p i is an integer, optionally wherein p i is multiplied by (1+ε) for an ε of small absolute value.

16. The system of claim 11 , wherein the number of parity symbols is set so that when l i (G) fraction of the parity symbols are lost, the number of received parity symbols of P[i] (i.e., p i R ) is approximately equal to (a) the number of missing symbols Γ[i] plus (b) the number of missing symbols of γ[i−τ] less the number of received symbols of G[i−τ], where this subtraction is bounded below by 0.

17. The system of claim 5 , wherein the parity symbols are set with the intention that the number of symbols of P[i] is sufficient to recover Γ[i] during time slot i (e.g.,

⌈

⁢

(

)

⌉

)

in the event that losses do not exceed a predetermined amount, optionally wherein the predetermined amount is defined as only losses as partial guard spaces.

18. The system of claim 5 , wherein the parity symbols are set with the intention that the number of symbols of G[i] is sufficient to recover γ[i] during time slot i (e.g.,

⌈

⁢

(

)

⌉

)

in the event that losses do not exceed a predetermined amount, optionally wherein the predetermined amount is defined as only losses as partial guard spaces.

19. The system of claim 5 , wherein some of the parity symbols of a data frame can be used to recover either the first or second component of the same data frame whereas other parity symbols can only be used to recover the first component of the data frame but not the second component, optionally wherein said some of the parity symbols comprise parity symbols of G[i] and wherein said other parity symbols comprise parity symbols of P[i].

20. The system of claim 5 , wherein some of the parity symbols of a data frame can be used to recover the first component but not the second component of the data frame and other parity symbols can be used to recover the second component but not the first component, optionally wherein said some of the parity symbols comprise parity symbols of P[i] and wherein said other parity symbols comprise parity symbols of G[i].

21. The system of claim 5 , wherein the number of parity symbols to be sent with the data of data frame i is set with the intention of (a) loss recovery during time slot i if there is no partial burst and losses defined as partial guard space and (b) if there is a partial burst starting in frame j encompassing frame i, (i) loss recovery by time slot (j+T) of the first component of frames j through (j+b j −1) (i.e., Γ[j:j+b j −1]), and (ii) loss recovery by time slot (i+τ) of the second component of frame i (i.e., γ[i]).

22. The system of claim 21 , wherein loss recovery of the first component of frames j through (j+b j −1) (i.e., Γ[j:j+b j −1]) occurs by time slot (j+τ−1).

23. The system of claim 21 , wherein the second component of the frames of the guard space will also be recovered in step (i), optionally wherein γ[j+b j :j+τ−1] and Γ[j:j+τ−1] are recovered by time slot (j+τ−1).

24. The system of claim 1 , wherein parity symbols are allocated such that the amount of parity symbols is minimized subject to predetermined performance targets for loss recovery.

25. The system of claim 1 , wherein the frame splitter splits the data frames into variable size components.

26. The system of claim 1 , wherein the frame splitter splits the data frames into fixed size components.

27. The system of claim 1 , wherein at least one of (a) splitting of the data frames into the two or more components or (b) allocating parity symbols for the two or more components or (c) packetizing the components and the parity symbols is based on feedback from a receiver.

28. The system of claim 1 , wherein at least one of (a) splitting of the data frames into the two or more components or (b) allocating parity symbols for the two or more components or (c) packetizing the components and the parity symbols is based on predictive analytics.

29. The system of claim 1 , wherein at least one of (a) splitting of the data frames into the two or more components or (b) allocating parity symbols for the two or more components or (c) packetizing the components and the parity symbols is based on machine learning.

30. The system of claim 1 , wherein at least one of (a) splitting of the data frames into the two or more components or (b) allocating parity symbols for the two or more components or (c) packetizing the components and the parity symbols is based on reinforcement learning.

31. The system of claim 1 , wherein the frame splitter is configured to select between splitting a given data frame into a single component or into two or more components, optionally wherein the frame splitter is configured to select how to split frames based on at least one of a frame-by-frame basis, a per-call basis, feedback from a receiver, predictive analytics, or machine learning.

32. The system of claim 1 , wherein at least one of the frame splitting or the parity symbol allocation is based on a heuristic.

33. The system of claim 1 , wherein parity symbol allocation includes at least one failsafe.

34. The system of claim 1 , wherein in the frame splitting encoder system is a compression-aware frame splitting encoder system that uses compression information from a frame compressor to determine FEC parameters for transmitting compressed frames.

35. The system of claim 34 , wherein the compression information comprises metadata indicating whether certain symbols are supplementary such that the data frame is useful without them but even better with them, optionally wherein such supplementary symbols are placed in γ[i] so that the non-supplementary symbols fit into Γ[i] to be recovered sooner.

36. The system of claim 34 , wherein the FEC parameters include at least one of frame splitting parameters, parity symbol allocation parameters, and/or packetization parameters.

37. The system of claim 34 , wherein the compression-aware FEC frame splitting encoder system is trained using machine learning to determine the FEC parameters for transmitting compressed frames, optionally wherein the machine learning is reinforcement learning.

38. The system of claim 34 , further comprising the frame compressor.

39. The system of claim 38 , wherein at least one mechanism within the compression-aware frame splitting encoder system and at least one mechanism within the frame compressor are trained jointly.

40. The system of claim 39 , wherein all considered mechanisms of the compression-aware frame splitting encoder system and the frame compressor are trained jointly.

41. The system of claim 39 , wherein the compression-aware frame splitting encoder system and the frame compressor are trained by alternating (a) fixing some of the mechanisms, and (b) training the non-fixed mechanisms.

42. The system of claim 1 , wherein packetizing the components and the parity symbols comprises dividing each component and each type of parity symbols into pieces and distributing the pieces across multiple packets, optionally wherein the pieces are equal size pieces and/or wherein packetization involves striping.

43. The system of claim 1 , further comprising an FEC-aware frame compressor that utilizes information from the frame splitting encoder system about FEC methodology and/or parameters to control compression of data into one or more compressed frames for the frame splitter.

44. The system of claim 43 , wherein the FEC-aware frame compressor performs selective compression based on anticipated parity allocation.

45. The system of claim 43 , wherein the information from the frame splitting encoder system includes at least one of frame splitting and/or parity symbol allocation information and/or indicators of frame splitting and/or parity symbol allocation information for future frames such as parameters of partial bursts and/or guardspaces.

46. The system of claim 43 , wherein the frame compressor selectively spreads information for a time slot over one or more additional time slots.

47. The system of claim 46 , wherein selectively spreading information comprises producing a lower resolution compression for an initial decompression and producing a higher resolution compression for a subsequent decompression.

48. The system of claim 46 , wherein selectively spreading information comprises creating a first compression to provide a less refined version of the data for the time slot and sending extra information during one or more later time slots to refine the prior information.

49. The system of claim 43 , wherein the FEC-aware frame compressor is trained using machine learning, optionally wherein the machine learning is reinforcement learning.

50. The system of claim 43 , wherein both the frame splitting encoder system and the FEC-aware frame compressor are trained together using machine learning, optionally wherein the machine learning is reinforcement learning.

51. The system of claim 43 , wherein the FEC-aware frame compressor utilizes information about FEC methodology and/or parameters to control the target size of the compressed frame for the FEC encoder.

52. The system of claim 51 , wherein the FEC-aware frame compressor is trained using machine learning, optionally wherein the machine learning is reinforcement learning where the reward is the number of symbols that are transmitted.

53. The system of claim 51 , wherein both the frame splitting encoder system and the FEC-aware frame compressor are trained together using machine learning, optionally wherein the machine learning is reinforcement learning.

54. The system of claim 51 , wherein the frame compressor tunes the target size based on the relationship between the size of the compressed frame and the number of parity symbols sent, optionally, wherein, the number of parity symbols sent is monotonically non-decreasing with the size of the compressed frame, so the marginal cost (in bandwidth usage) of increasing the target size may differ based on the state of the system and target size.

55. The system of claim 51 , wherein the frame compressor tunes the amount of data sent per frame without spreading information content, optionally by choosing to compress to a smaller number of symbols (providing less information/a lower resolution frame) for a frame i and then for the next frame i+1 compressing to more symbols (providing more information about the next frame/higher resolution).

56. The system of claim 51 , wherein the frame compressor tunes the amount of data sent per frame and also spreads information content, optionally by choosing to compress to a smaller number of symbols (providing less information/a lower resolution frame) for a frame i and then for the next frame i+1 (a) compressing to more symbols (providing more information about the next frame/higher resolution) while also (b) sending information to help lead to a better resolution of frame i.

57. The system of claim 51 , wherein the frame compressor selectively spreads information for a time slot over one or more additional time slots, optionally by creating a first compression to provide a less refined version of the data for the time slot and sending extra information during one or more later time slots to refine the prior information.

58. The system of claim 51 , wherein the FEC parameters include at least one of frame splitting parameters, parity symbol allocation parameters, and/or packetization parameters.

59. The system of claim 51 , wherein the frame compressor creates one or more parts as part of compression (and may or may not include metadata about the parts) and distributes the parts over one or more compressed frames, optionally wherein the frame compressor makes two parts, allocates the first part to the current compressed frame and the second part to the subsequent compressed frame, and then uses metadata to indicate that the FEC scheme should place the second part in the first component to lead to loss recovery at a tolerable latency.

60. The system of claim 51 , wherein the frame compressor adjusts the sizes of compressed frames by reducing or increasing the video quality for compressed frames to better control the amount of data needed in compressed frames such that the displayed frames may reduce the granularity of certain frames if necessary to keep the video consistent with the understanding that having a higher quality frame available than is displayed is still useful (e.g., for better inter-frame compression of future frames).

61. The system of claim 1 embodied as a sender device that implements the frame splitter, the parity symbol generator, and the packetizer.

62. The system of claim 1 embodied as a computer program product comprising a tangible, non-transitory computer readable medium having embodied therein program code which, when executed by at least one processor, implements the frame splitter, the parity symbol generator, and the packetizer.

63. The system of claim 1 embodied as an integrated circuit having circuitry configured to implement the frame splitter, the parity symbol generator, and the packetizer.

64. A receiver device that decodes lost data and/or parity symbols in accordance with the encoding of claim 1 .

65. A frame splitting encoder method comprising:

splitting each of a number of data frames i into a plurality of components including at least a first component Γ[i] and a second component γ[i];

allocating parity symbols for the components to ensure that (a) if the fraction of packets lost during time slot i is at most l i (G) and the data of all prior frames is available, then the parity symbols sent during a time slot suffice to recover the lost data during that time slot, and (b) if time slot i is part of a partial burst starting in time slot j (i.e., for each time slot z∈{j, . . . , j+b j −1}, l z or fewer fraction of the packets sent during the time slot are lost where i∈{j, . . . , j+b j −1}) followed by a partial guard space (i.e., for each time slot z∈{j+b j , . . . , j+b j −1+τ}, l z (G) or fewer fraction of the packets sent during the time slot are lost) and all the data of frames before the start of the partial burst is available, then the data for frame i is recovered within τ time slots; and

packetizing the components and the parity symbols.

66. The method of claim 65 , wherein parity symbols are allocated such that all lost data for the first component of the frames is to be recovered by (τ−1) time slots after the start of the partial burst and the second component of each frame of the partial burst is recovered τ time slots later.

67. The method of claim 66 , wherein the second component of the frames in a partial guard space after the partial burst are recovered by (τ−1) time slots after the start of the partial burst excluding future frames τ or more time slots after the start of the partial burst.

68. The method of claim 65 , wherein parity symbols are allocated such that there are two or more types of parity symbols where one type of parity symbol for a frame is independent of the symbols of the second component of the same frame.

69. The method of claim 68 , wherein the parity symbols allocated for a data frame i include at least a first set of parity symbols P[i] and a second set of parity symbols G[i] based on the plurality of components, wherein the second set of parity symbols is configured to ensure that (a) some parity symbols of a data frame can be used to recover the first component of the same data frame, (b) some parity symbols of a data frame can be used to recover the second component of the same data frame, and (c) some parity symbols of a data frame cannot be used to recover the second component of the data frame.

70. The method of claim 69 , wherein the number of parity symbols allocated to P[i] and G[i] is based on the parameter l i (G) .

71. The method of claim 70 , wherein the parameter l i (G) is set to approximate an upper bound on the fraction of packets that may be lost during time slot i if there is no partial burst encompassing time slot i for which loss recovery is needed.

72. The method of claim 70 , wherein l i (G) is set so that the fraction of losses during time slot i is no more than l i (G) with some probability.

73. The method of claim 69 , wherein allocating the parity symbols comprises a two-stage parity allocation of (a) pre-allocating p i+τ during time slot i for robustness to partial bursts then (b) increasing the size of p i during time slot i for robustness to loss in the partial guard space; the size of G[i] can then be set during time slot i.

74. The method of claim 69 , wherein the frame splitting encoder method is multimodal.

75. The method of claim 69 , wherein the number of parity symbols to be sent with the data of data frame i is set with the intention of (a) loss recovery during time slot i if there is no partial burst (and losses defined as partial guard space) and (b) loss recovery during time slot i of the second component of frame (i−τ) (i.e., γ[i−τ]) which reflects if the partial burst encompasses frame (i−τ).

76. The method of claim 75 , wherein:

if the partial burst includes time slot i, then some of the parity symbols are used to recover a subset of Γ[i] and/or some of the parity symbols are used to recover Γ[i−τ+1:i−1]; and

77. The method of claim 75 , wherein the number of parity symbols p i +g i is set so that when l i (G) fraction of the parity symbols are lost and a partial burst encompasses time slot (i−τ), the number of received parity symbols p i R +g i R is approximately equal to (a) the number of missing symbols Γ[i] plus (b) the number of missing symbols of γ[i−τ] less the number of received symbols of G[i−τ], where this subtraction is bounded below by 0.

78. The method of claim 77 , wherein g i is set so that the at least g i R =(1−l i (G) )g i received symbols of G[i] suffice to recover the lost symbols of γ[i], leaving pf symbols of P[i] to recover the lost symbols of Γ[i] and γ[i−τ].

79. The method of claim 77 , wherein p i is set approximately as follows:

p i (1− l i (G) )=γ i l i (G) +max(0, l i−τ υ i−τ −(1− l i−τ ) g i−τ ),

leading to:

⌈

⁢

(

)

max

⁡

(

⁢

(

)

⁢

)

(

)

⌉

and the ceiling is taken to ensure p i is an integer, optionally wherein p i is multiplied by (1+ε) for an ε of small absolute value.

80. The method of claim 75 , wherein the number of parity symbols is set so that when l i (G) fraction of the parity symbols are lost, the number of received parity symbols of P[i] (i.e., p i R ) is approximately equal to (a) the number of missing symbols Γ[i] plus (b) the number of missing symbols of γ[i−τ] less the number of received symbols of G[i−τ], where this subtraction is bounded below by 0.

81. The method of claim 69 , wherein the parity symbols are set with the intention that the number of symbols of P[i] is sufficient to recover Γ[i] during time slot i (e.g.,

(

⌈

⁢

(

)

⌉

)

in the event that losses do not exceed a predetermined amount, optionally wherein the predetermined amount is defined as only losses as partial guard spaces.

82. The method of claim 69 , wherein the parity symbols are set with the intention that the number of symbols of G[i] is sufficient to recover γ[i] during time slot i (e.g.,

(

⌈

⁢

(

)

⌉

)

in the event that losses do not exceed a predetermined amount, optionally wherein the predetermined amount is defined as only losses as partial guard spaces.

83. The method of claim 69 , wherein some of the parity symbols of a data frame can be used to recover either the first or second component of the same data frame whereas other parity symbols can only be used to recover the first component of the data frame but not the second component, optionally wherein said some of the parity symbols comprise parity symbols of G[i] and wherein said other parity symbols comprise parity symbols of P[i].

84. The method of claim 69 , wherein some of the parity symbols of a data frame can be used to recover the first component but not the second component of the data frame and other parity symbols can be used to recover the second component but not the first component, optionally wherein said some of the parity symbols comprise parity symbols of P[i] and wherein said other parity symbols comprise parity symbols of G[i].

85. The method of claim 69 , wherein the number of parity symbols to be sent with the data of data frame i is set with the intention of (a) loss recovery during time slot i if there is no partial burst and losses defined as partial guard space and (b) if there is a partial burst starting in frame j encompassing frame i, (i) loss recovery by time slot (j+τ) of the first component of frames j through (j+b j −1) (i.e., Γ[j:j+b j −1]), and (ii) loss recovery by time slot (i+τ) of the second component of frame i (i.e., γ[i]).

86. The method of claim 85 , wherein loss recovery of the first component of frames j through (j+b j −1) (i.e., Γ[j:j+b j −1]) occurs by time slot (j+τ−1).

87. The method of claim 85 , wherein the second component of the frames of the guard space will also be recovered in step (i), optionally wherein γ[j+b j :j+τ−1] and Γ[j:j+τ−1] are recovered by time slot (j+τ−1).

88. The method of claim 65 , wherein parity symbols are allocated such that the amount of parity symbols is minimized subject to predetermined performance targets for loss recovery.

89. The method of claim 65 , wherein the data frames are split into variable size components.

90. The method of claim 65 , wherein the data frames are split into fixed size components.

91. The method of claim 65 , wherein at least one of (a) splitting of the data frames into the two or more components or (b) allocating parity symbols for the two or more components or (c) packetizing the components and the parity symbols is based on feedback from a receiver.

92. The method of claim 65 , wherein at least one of (a) splitting of the data frames into the two or more components or (b) allocating parity symbols for the two or more components or (c) packetizing the components and the parity symbols is based on predictive analytics.

93. The method of claim 65 , wherein at least one of (a) splitting of the data frames into the two or more components or (b) allocating parity symbols for the two or more components or (c) packetizing the components and the parity symbols is based on machine learning.

94. The method of claim 65 , wherein at least one of (a) splitting of the data frames into the two or more components or (b) allocating parity symbols for the two or more components or (c) packetizing the components and the parity symbols is based on reinforcement learning.

95. The method of claim 65 , wherein splitting each of a number of data frames i into a plurality of components comprises selecting between splitting a given data frame into a single component or into two or more components.

96. The method of claim 65 , wherein at least one of the frame splitting or the parity symbol allocation is based on a heuristic.

97. The method of claim 65 , wherein parity symbol allocation includes at least one failsafe.

98. The method of claim 65 , wherein the frame splitting encoder method is a compression-aware frame splitting encoder method that uses compression information from a frame compressor to determine FEC parameters for transmitting compressed frames.

99. The method of claim 98 , wherein the compression information comprises metadata indicating whether certain symbols are supplementary such that the data frame is useful without them but even better with them, optionally wherein such supplementary symbols are placed in γ[i] so that the non-supplementary symbols fit into Γ[i] to be recovered sooner.

100. The method of claim 98 , wherein the FEC parameters include at least one of frame splitting parameters, parity symbol allocation parameters, and/or packetization parameters.

101. The method of claim 98 , wherein the compression-aware frame splitting encoder method is trained using machine learning to determine the FEC parameters for transmitting compressed frames, optionally wherein the machine learning is reinforcement learning.

102. The method of claim 98 , further comprising the frame compressor producing the compression information.

103. The method of claim 102 , wherein at least one mechanism within the compression-aware frame splitting encoder method and at least one mechanism within the frame compressor are trained jointly.

104. The method of claim 103 , wherein all considered mechanisms of the compression-aware PEC frame splitting encoder method and the frame compressor are trained jointly.

105. The method of claim 103 , wherein the compression-aware frame splitting encoder method and the frame compressor are trained by alternating (a) fixing some of the mechanisms, and (b) training the non-fixed mechanisms.

106. The method of claim 65 , wherein packetizing the components and the parity symbols comprises dividing each component and each type of parity symbols into pieces and distributing the pieces across multiple packets, optionally wherein the pieces are equal size pieces and/or wherein packetization involves striping.

107. The method of claim 65 , further comprising utilizing information from the frame splitting encoder method about FEC methodology and/or parameters to control compression of data into one or more compressed frames for the frame splitter.

108. The method of claim 107 , wherein controlling compression performs selective compression based on anticipated parity allocation.

109. The method of claim 107 , wherein the information from the frame splitting encoder method includes at least one of frame splitting and/or parity symbol allocation information and/or indicators of frame splitting and/or parity symbol allocation information for future frames such as parameters of partial bursts and/or guardspaces.

110. The method of claim 107 , wherein controlling compression selectively spreads information for a time slot over one or more additional time slots.

111. The method of claim 110 , wherein selectively spreading information comprises producing a lower resolution compression for an initial decompression and producing a higher resolution compression for a subsequent decompression.

112. The method of claim 110 , wherein selectively spreading information comprises creating a first compression to provide a less refined version of the data for the time slot and sending extra information during one or more later time slots to refine the prior information.

113. The method of claim 107 , wherein the FEC-aware frame compressor is trained using machine learning, optionally wherein the machine learning is reinforcement learning.

114. The method of claim 107 , wherein both the frame splitting encoder method and the FEC-aware frame compressor are trained together using machine learning, optionally wherein the machine learning is reinforcement learning.

115. The method of claim 107 , wherein the FEC-aware frame compressor utilizes information about FEC methodology and/or parameters to control the target size of the compressed frame for the FEC encoder.

116. The method of claim 115 , wherein the FEC-aware frame compressor is trained using machine learning, optionally wherein the machine learning is reinforcement learning where the reward is the number of symbols that are transmitted.

117. The method of claim 115 , wherein both the frame splitting encoder method and the FEC-aware frame compressor are trained together using machine learning, optionally wherein the machine learning is reinforcement learning.

118. The method of claim 115 , wherein the frame compressor tunes the target size based on the relationship between the size of the compressed frame and the number of parity symbols sent, optionally, wherein the number of parity symbols sent is monotonically non-decreasing with the size of the compressed frame, so the marginal cost (in bandwidth usage) of increasing the target size may differ based on the state of the system and target size.

119. The method of claim 115 , wherein the frame compressor tunes the amount of data sent per frame without spreading information content, optionally by choosing to compress to a smaller number of symbols (providing less information/a lower resolution frame) for a frame i and then for the next frame i+1 compressing to more symbols (providing more information about the next frame/higher resolution).

120. The method of claim 115 , wherein the frame compressor tunes the amount of data sent per frame and also spreads information content, optionally by choosing to compress to a smaller number of symbols (providing less information/a lower resolution frame) for a frame i and then for the next frame i+1 (a) compressing to more symbols (providing more information about the next frame/higher resolution) while also (b) sending information to help lead to a better resolution of frame i.

121. The method of claim 115 , wherein the frame compressor selectively spreads information for a time slot over one or more additional time slots, optionally by creating a first compression to provide a less refined version of the data for the time slot and sending extra information during one or more later time slots to refine the prior information.

122. The method of claim 115 , wherein the FEC parameters include at least one of frame splitting parameters, parity symbol allocation parameters, and/or packetization parameters.

123. The method of claim 115 , wherein the frame compressor creates one or more parts as part of compression (and may or may not include metadata about the parts) and distributes the parts over one or more compressed frames, optionally wherein the frame compressor makes two parts, allocates the first part to the current compressed frame and the second part to the subsequent compressed frame, and then uses metadata to indicate that the FEC scheme should place the second part in the first component to lead to loss recovery at a tolerable latency.

124. The method of claim 115 , wherein the frame compressor is configured to adjust the sizes of compressed frames by reducing or increasing the video quality for compressed frames to control the amount of data needed in compressed frames such that the frame compressor may reduce the granularity of certain displayed frames if necessary to keep the video consistent, optionally wherein a higher quality frame than is displayed is available to the frame compressor for inter-frame compression.

Assignments (1)

ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Nov 11, 2025

From: RUDOW, MICHAEL H.

To: TRESEDER AI, INC.

Reel/Frame 072861/0802 →

Continuity (2)

Continuation PCTUS2024043054 · Aug 20, 2024

Provisional Application 63655861 · Jun 4, 2024

References Cited (137)

US 5699365A · Klayman et al. · 1997 [cited by applicant]

US 6421387B1 · Rhee · 2002 [cited by examiner]

US 6694478B1 · Martinian et al. · 2004 [cited by applicant]

US 7257664B2 · Zhang · 2007 [cited by applicant]

US 8352832B2 · Khisti et al. · 2013 [cited by applicant]

US 8375266B2 · Zhang · 2013 [cited by applicant]

US 8775889B2 · Zhang · 2014 [cited by applicant]

US 9209897B2 · Amitai et al. · 2015 [cited by applicant]

US 9641803B1 · Badr et al. · 2017 [cited by applicant]

US 9843413B2 · Badr et al. · 2017 [cited by applicant]

US 10833710B2 · Caramma · 2020 [cited by applicant]

US 10979175B2 · Low et al. · 2021 [cited by applicant]

US 11036525B2 · Momchilov · 2021 [cited by applicant]

US 11303690B1 · Bhattacharyya · 2022 [cited by examiner]

US 20130039410A1 · Tan · 2013 [cited by examiner]

US 20130070844A1 · Malladi et al. · 2013 [cited by applicant]

US 20130097470A1 · Hwang et al. · 2013 [cited by applicant]

US 20130156420A1 · Amitai et al. · 2013 [cited by applicant]

US 20170279558A1 · Badr et al. · 2017 [cited by applicant]

US 20180034583A1 · Low · 2018 [cited by examiner]

US 20190007069A1 · Caramma · 2019 [cited by applicant]

US 20190339997A1 · Momchilov · 2019 [cited by applicant]

US 20200044772A1 · Low et al. · 2020 [cited by applicant]

US 20220124543A1 · Orhan et al. · 2022 [cited by applicant]

US 20230106959A1 · Ananthanarayanan et al. · 2023 [cited by applicant]

WO 2019213556A1 · 2019 [cited by applicant]

Rudow et al., Streaming codes for variable-size messages, IEEE Trans. on Information Theory, vol. 68, No. 9, pp. 5823-5849. (Year: 2022). [cited by examiner]

Ellis et al., Performance analysis of AL-FEC for RTP-based streaming video traffic to residential users, IEEE, pp. 1 to 6. (Year: 2012). [cited by examiner]

International Search Report and Written Opinion for International Application No. PCT/US2024/043054, mailed Dec. 31, 2024 (19 pages). [cited by applicant]

Lin, C., “A RED-FEC Mechanism for Video TransmissionOver WLANs,” IEEE Transactions on Broadcasting, vol. 54, No. 3, Sep. 2008 (8 pages). [cited by applicant]

Lykouris, T., et al., “Competitive Caching with Machine Learned Advice,” 2021. J. ACM 68, 4, Article 24 (Jul. 2021), (pp. 24:1-24:25 (25 pages). [cited by applicant]

Ma, S., et al., “Image and V+B63:B75ideo Compression with Neural Networks: A Review,” IEEE Transactions On Circuits And Systems For Video Technology, arXiv:1904.03567v2 [cs.CV] Apr. 10, 2019 (16 pages). [cited by applicant]

Martinian, E., et al., “Burst Erasure Correction Codes With Low Decoding Delay,” IEEE Transactions On Information Theory, vol. 50, No. 10, Oct. 2004 pp. 2494-2502 (9 pages). [cited by applicant]

Maturana, F., et al., “Access-optimal Linear MDS Convertible Codes for All Parameters,” available on arXiv, 2020 (6 pages). [cited by applicant]

Mazyavkina, N., et al., “Reinforcement learning for combinatorial optimization: A survey,” Computers & Operations Research 134 (2021) 105400 (15 pages). [cited by applicant]

McCanne, S. et al., “Joint Source/Channel Coding for Multicast Packet Video,” Proceedings., International Conference on Image Processing, Washington, DC, USA, 1995, pp. 25-28 vol. 1, doi: 10.1109/ICIP.1995.529030 (4 pag… [cited by applicant]

Mitzenmacher, M., “A Model for Learned Bloom Filters, and Optimizing by Sandwiching,” 32nd Conference on Neural Information Processing Systems (NeurIPS 2018), Montréal, Canada (10 Pages). [cited by applicant]

Moerland, T., et al., “Model-based Reinforcement Learning: A Survey,” Foundations and Trends® in Machine Learning, vol. 16, No. 1, pp. 1-118, 2023 (43 pages). [cited by applicant]

Nagy, M., et al., “Congestion Control using FEC for Conversational Multimedia Communication,” MMSys '14 Mar. 19-21, 2014, Singapore, Singapore (12pages). [cited by applicant]

Narra, H., et al., “Collage Inference: Using Coded Redundancy for Lowering Latency Variation in Distributed Image Classification Systems,” 2020 IEEE 40th International Conference on Distributed Computing Systems (ICDCS)… [cited by applicant]

Nichols, K., et al., “Controlled Delay Active Queue Management,” Internet Engineering Task Force (IETF), Jan. 2018 (25 pages). [cited by applicant]

Nowroozi, E., et al., “A survey of machine learning techniques in adversarial image forensics,” Computers & Security 100 (2021) 102092 (25 pages). [cited by applicant]

Orosz, P., et al., “A Case Study on Correlating Video QoS and QoE,” 2014 IEEE Network Operations and Management Symposium (NOMS) (5 pages). [cited by applicant]

O'Shea, T. et al., “An Introduction to Deep Learning for the Physical Layer,” IEEE Transactions On Cognitive Communications And Networking, vol. 3, No. 4, Dec. 2017. [cited by applicant]

Pan, R., et al., “Proportional Integral Controller Enhanced (PIE): A Lightweight Control Scheme to Address the Bufferbloat Problem,” Internet Engineering Task Force (IETF), Feb. 2017 (30 pages). [cited by applicant]

Park, K., “AFEC: An Adaptive Forward Error-Correction Protocol and Its Analysis,” Department of Computer Science Technical Reports, Report No. 97-038, 1997 (25 pages). [cited by applicant]

Perkins, C., et al., “A Survey of Packet Loss Recovery Techniques for Streaming Audio,” IEEE Network Sep./Oct. 1998 pp. 40-48 (9 pages). [cited by applicant]

Ponlatha, S., et al., “Comparison of Video Compression Standards,” International Journal of Computer and Electrical Engineering, vol. 5, No. 6, Dec. 2013 (7 pages). [cited by applicant]

Powell, W., “A unified framework for stochastic optimization,” European Journal of Operational Research (2018) pp. 1-27 (27 pages). [cited by applicant]

Raghavendra, R., et al., “Characterizing High-bandwidth Real-time Video Traffic in Residential Broadband Networks,” WiOpt'10: Modeling and Optimization in Mobile, Ad Hoc, and Wireless Networks, May 2010, Avignon, France… [cited by applicant]

Rao, N., et al., “Analysis of the Effect of QoS on Video Conferencing QoE,” https://ieeexplore.ieee.org/abstract/document/8766591, 2019 (6 pages). [cited by applicant]

Rippel, O., et al., “Learned Video Compression,” Computer Vision Foundation, arXiv:1811.06981, Nov. 16, 2018 (10 pages). [cited by applicant]

Rosenberg, J., et al., “An RTP Payload Format for Generic Forward Error Correction,” Network Working Group, Columbia University, Dec. 1999 (26 pages). [cited by applicant]

Rudow, M. et al., “A locality-based approach for coded computation,” arXiv:2002.02440v1 [cs.IT], Feb. 6, 2020 (18 pages). [cited by applicant]

Rudow, M. et al., “A locality-based lens for Coded computation,” retrieved from the internet at: //2021 IEEE International Symposium on Information Theory (ISIT) © 2021 IEEE | DOI: 10.1109/ISIT45174.2021.9518056// (6 pa… [cited by applicant]

Rudow, M., et al., “Compression-informed coded computing,” In 2023 IEEE International Symposium on Information Theory (ISIT), pp. 2177-2182, 2023 (6 pages). [cited by applicant]

Rudow, M., “Discrete Logarithm and Minimum Circuit Size,” Inf. Process. Lett. (2017) http://dx.doi.org/10.1016/j.ipl.2017.07.005 (10 pages). [cited by applicant]

Rudow, M. et a., “Learning-Augmented Streaming Codes are Approximately Optimal for Variable-Size Messages,” arXiv:2205.08521v1 [cs.IT] Extended Version, May 17, 2022 (13 pages). [cited by applicant]

Rudow, M., et al., “Learning-augmented streaming codes for variable-size messages under partial burst losses,” in 2023 IEEE International Symposium on Information Theory (ISIT), 2023, (20 pages). [cited by applicant]

Rudow, M., et al., “On expanding the toolkit of locality-based coded computation to the coordinates of inputs,” 2023 IEEE International Symposium on Information Theory (ISIT) (10 pages). [cited by applicant]

Rudow, M., et al., “Learning-augmented streaming codes are approximately optimal for variable-size messages,” 2022 IEEE International Symposium on Information Theory (ISIT). IEEE, pp. 474-479, 2022 (6 pages). [cited by applicant]

Rudow, M., et al., “Learning-Augmented Streaming Codes For Variable-Size Messages Under Partial Burst Losses,” In 2023 IEEE International Symposium on Information Theory (ISIT), pp. 1101-1106, 2023 (6 pages). [cited by applicant]

Rudow, M., et al., “Online Versus Offline Rate in Streaming Codes for Variable-Size Messages,” arXiv:2006.03045v2 [cs.IT] Feb. 27, 2023 (21 pages). [cited by applicant]

Rudow, M., et al., “Online Versus Offline Rate In Streaming Codes For Variable-Size Messages,” 2020 IEEE International Symposium on Information Theory (ISIT), 2020 (6 pages). [cited by applicant]

Rudow, M., et al., “Online versus offline rate in streaming codes for variable-size messages,” IEEE Transactions on Information Theory, vol. 69, No. 6, pp. 3674-3690, 2023 (17 pages). [cited by applicant]

Rudow, M., et al., “Streaming codes for variable-size arrivals,” 2018 56th Annual Allerton Conference on Communication, Control, and Computing, (Allerton). IEEE, 2018 (8 pages). [cited by applicant]

Rudow, M., et al., “Streaming Codes for Variable-Size Messages,” IEEE Transactions on Information Theory, vol. 68, No. 9, pp. 5823-5849, Sep. 2022 (27 pages). [cited by applicant]

Rudow, M., et al., “Tambur: Efficient loss recovery for videoconferencing via streaming codes,” In Proceedings of the 20th USENIX Symposium on Networked Systems Design and Implementation, Apr. 17-19, 2023 (20 pages). [cited by applicant]

Rudow, M., “Efficient loss recovery for videoconferencing via streaming codes and machine learning,” Carnegie Mellon University, Thesis, retrieved from the internet at //https://doi.org/10.1184/R1/24992973.v1//, May 202… [cited by applicant]

Sharma, S., “Active Queue Management for Forward Error Correction,” International Journal of Computing and Business Research (IJCBR) ISSN (Online) : 2229-6166, vol. 3, Issue 2, May 2012 (7 pages). [cited by applicant]

Sztrik, J., “Basic Queueing Theory,” https://irh.inf.unideb.hu/˜jsztrik/education/16/SOR_Main_Angol.pdf, 2016 (246 pages). [cited by applicant]

Sullivan, G., et al., “Video Compression-From Concepts to the H.264/AVC Standard,” Proceedings of the IEEE, vol. 93, No. 1, Jan. 2005 (14 pages). [cited by applicant]

Tan, W., et al., “Video Multicast using Layered FEC and Scalable Compression,” in IEEE Transactions on Circuits and Systems for Video Technology, vol. 11, No. 3, pp. 373-386, Mar. 2001, doi: 10.1109/76.911162 (30 pages). [cited by applicant]

Uehara, M. et al., “A Review of Off-Policy Evaluation in Reinforcement Learning,” arXiv:2212.06355v1 [stat.ML] Dec. 13, 2022 (27 pages). [cited by applicant]

Usman, M., et al., “Survey of Error Concealment Techniques: Research Directions and Open Issues,” 2015 Picture Coding Symposium (PCS), Cairns, QLD, Australia, 2015, pp. 233-238, doi: 10.1109/PCS.2015.7170081 (7 pages). [cited by applicant]

Wah, B., et al. “Survey of Error-Concealment Schemes for Real-Time Audio and Video Transmissions over the Internet,” Proceedings International Symposium on Multimedia Software Engineering 2000 (8 pages). [cited by applicant]

Wang, Y., et al., “Error Control and Concealment for Video Communication: A Review,” Proceedings of the IEEE, vol. 86, No. 5, May 1998 (24 pages). [cited by applicant]

Watson, M. et al., “Forward Error Correction (FEC) Framework,” Internet Engineering Task Force (IETF), Oct. 2011 (42 pages). [cited by applicant]

Adams, R., “Active Queue Management: A Survey,” IEEE Communications Surveys & Tutorials, vol. 15, No. 3, Third Quarter 2013 (52 pages). [cited by applicant]

Adan, I., et al., “Queueing Theory,” retrieved from the internet at https://citeseerx.ist.psu.edu/document?repid=rep1&type=pdf&doi=097d866985fc04b986d1f28885de7a0c1e89fce9, Feb. 14, 2001 (180 pages). [cited by applicant]

Adler, N., et al., “Burst-Erasure Correcting Codes With Optimal Average Delay,” IEEE Transactions On Information Theory, vol. 63, No. 5, May 2017, pp. 2848-2865 (18 pages). [cited by applicant]

Agrawal, A., et al., “A Rewriting System for Convex Optimization Problems,” arXiv:1709.04494v2 [math.OC] Jan. 22, 2019, (18 pages). [cited by applicant]

Alemu,T., “The Interaction of Forward Error Correction and Active Queue Management,” JDIR'04 : 6èmes Journées Doctorales Informatique et Réseau, Nov. 2004, Lannion, France. lirmm-00108649. [cited by applicant]

Almomani, O., et al., “Impact of Large Block FEC with Different Queue Sizes of Drop Tail and RED Queue Policy on Video Streaming Quality over Internet,” 2010 Second International Conference on Network Applications, Prot… [cited by applicant]

Almomani, O., et al., “Performance Study of Large Block FEC with Drop Tail for Video Streaming over the Internet,” 2009 First International Conference on Networks & Communications (4 pages). [cited by applicant]

Alwahab, D., “A Simulation-Based Survey of Active Queue Management Algorithms,” ICCBN 2018, Feb. 24-26, 2018, Singapore, Singapore, DOI: https://doi.org/10.1145/3193092.3193106 (7 pages). [cited by applicant]

Arulkumaran, K., et al., “A Brief Survey of Deep Reinforcement Learning,” IEEE Signal Processing Magazine, Special Issue On Deep Learning For Image Understanding (Arxiv Extended Version), arXiv:1708.05866v2 [cs.LG] Sep.… [cited by applicant]

Author Unknown, “Series G: Transmission Systems and Media, Digital Systems and Networks, Quality of service and performance,” ITU-T, Telecommunication Standardization Sector of ITU, Sep. 2001 (18 pages). [cited by applicant]

Badr, A., et al., “Embedded MDS Codes for Multicast Streaming,” 2015 IEEE International Symposium on Information Theory (ISIT) (5 pages). [cited by applicant]

Baguda, Y., et al., “Adaptive FEC Error Control Scheme for Wireless video Transmission,” 2010 The 12th International Conference on Advanced Communication Technology (ICACT) (5 pages). [cited by applicant]

Bandung, Y., “QoS Analysis for WebRTC Videoconference on Bandwidth-Limited Network,” The 20th International Symposium on Wireless Personal Multimedia Communications (WPMC2017), 2017 IEEE pp. 547-553 (7 pages). [cited by applicant]

Boykov, Y., et al., “An Experimental Comparison of Min-Cut/Max-Flow Algorithms for Energy Minimization in Vision,” IEEE Transactions On Pattern Analysis And Machine Intelligence, vol. 26, No. 9, Sep. 2004, pp. 1124-1137… [cited by applicant]

Brockners, F. “The Case for FEC fueled TCP-like Congestion Control,” Kommunikation in Verteilten Systemen (KiVS) 11. ITG/GI-Fachtagung. Darmstadt, 2.-5. Mar. 1999, retrieved from the internet https://citeseerx.ist.psu.e… [cited by applicant]

Canese, L., et al., “Multi-Agent Reinforcement Learning: A Review of Challenges and Applications,” Appl. Sci. 2021, 11, 4948. https://doi.org/ 10.3390/app11114948 (25 pages). [cited by applicant]

Carlucci, G., et al., “Analysis and Design of the Google Congestion Control for Web Real-time Communication (WebRTC),” Proceedings of the 7th International Conference on Multimedia Systems, retrieved from the internet a… [cited by applicant]

Chang, H., et al., “Can You See Me Now? A Measurement Study of Zoom, Webex, and Meet,” IMC '21, Nov. 2-4, 2021, Virtual Event, USA, pp. 216-228 (13 pages). [cited by applicant]

Clarke, R., “Image and Video Compression: A Survey,” Creative Commons License, vol. 10, pp. 20-32, 1999 (13 pages). [cited by applicant]

Cohen, A., et al., “Adaptive Causal Network Coding With Feedback,” EEE Transactions On Communications, vol. 68, No. 7, Jul. 2020 (17 pages). [cited by applicant]

Dischinger, M., et al., “Characterizing Residential Broadband Networks,” 07, Oct. 24-26, 2007, San Diego, California, USA, pp. 43-56, (14 pages). [cited by applicant]

Domanovitz, D., “An Explicit Rate-Optimal Streaming Code for Channels With Burst and Arbitrary Erasures,” IEEE Transactions On Information Theory, vol. 68, No. 1, Jan. 2022 (19 pages). [cited by applicant]

Duanmu, et al., “A Quality-of-Experience Index for Streaming Video,” IEEE Journal Of Selected Topics In Signal Processing, vol. 11, No. 1, Feb. 2017 (13 pages). [cited by applicant]

Dudzicz, D., et al., “An Explicit Construction of Optimal Streaming Codes for Channels With Burst and Arbitrary Erasures,” IEEE Transactions On Communications, vol. 68, No. 1, Jan. 2020 pp. 12-25 (14 pages). [cited by applicant]

Falk, B., et al., “Properties of Constacyclic Codes Under the Schur Product,” arXiv:1810.07630v2 [cs.IT] Oct. 18, 2018 (24 pages). [cited by applicant]

Falk, B., et al., “Properties of Constacyclic Codes Under the Schur Product,” Designs, Codes and Cryptography (2020) 88:993-1021 (29 pages). [cited by applicant]

Floyd, S. et al., “Random Early Detection Gateways for Congestion Avoidance,” IEEE/ACM Transactions on Networking. vol. I. No. 1. Aug. 1993 (17 pages). [cited by applicant]

Forney, Jr., G., “Burst-Correcting Codes for the Classic Bursty Channel,” IEEE Transactions On Communications Technology, Oct. 1971, pp. 772-781 (10 pages). [cited by applicant]

Friedman, T., et al., “RTP Control Protocol Extended Reports (RTCP XR),” Network Working Group, Nov. 2003 (55 pages). [cited by applicant]

Frossard, P., “Joint Source/FEC Rate Selection for Quality-Optimal MPEG-2 Video Delivery,” IEEE Transactions On Image Processing, vol. 10, No. 12, Dec. 2001(11 pages). [cited by applicant]

Geist, M., et al., “Off-policy Learning With Eligibility Traces: A Survey,” Journal of Machine Learning Research 15 (2014) 289-333 (45 pages). [cited by applicant]

Gettys, J., “Bufferbloat: Dark Buffers in the Internet,” Published by the IEEE Computer Society, May/Jun. 2011 (2 pages). [cited by applicant]

Ghavamzadeh, M., et al., “Bayesian Reinforcement Learning: A Survey,” Foundations and Trends® in Machine Learning, vol. 8, No. 5-6 (2015) 359-483, 2015, 28 pages. [cited by applicant]

Gilbert, E., “Capacity of a Burst-Noise Channel,” The Bell System Technical Journal, Sep. 1960, pp. 1253-1265 (13 pages). [cited by applicant]

Ha, H., et al., “Layer-based RED-FEC (L-RED-FEC) method for wireless scalable video streaming,” Electronics Letters Sep. 25, 2014 vol. 50 No. 20 pp. 1438-1440 (2 pages). [cited by applicant]

Huo, Y., et al., “A Tutorial and Review on Inter-Layer FEC Coded Layered Video Streaming,” IEEE Communications Surveys & Tutorials 2015 (44 pages). [cited by applicant]

Ibrahim, I. et al., “Task Scheduling Algorithms in Cloud Computing: A Review,” Turkish Journal of Computer and Mathematics Education, vol. 12 No. 4 (2021), pp. 1041-1053 (13 pages). [cited by applicant]

Jeon, Y., et al., “Blind Detection for MIMO Systems With Low-Resolution ADCs Using Supervised Learning,” IEEE ICC 2017 Signal Processing for Communications Symposium (6 pages). [cited by applicant]

Jiang, P., et al., “Wireless Semantic Communications for Video Conferencing,” IEEE Journal on Selected Areas in Communications, vol. 41, No. 1, Jan. 2023 (15 pages). [cited by applicant]

Kaelbling, L., et al., “Reinforcement Learning: A Survey,” Journal of Articial Intelligence Research 4 (1996) 237-285 (49 pages). [cited by applicant]

Kazemi, M., et al., “A review of temporal video error concealment techniques and their suitability for HEVC and VVC,” Multimedia Tools and Applications (2021) 80:12685-12730 (46 pages). [cited by applicant]

Kotz, D., et al., “Crawdad: A Community, Resource for Archiving, Wireless Data at Dartmouth,” Conferences—HP Labs, www.computer.org/pervasive (3 pages). [cited by applicant]

Krishan, M., et al., “A Quadratic Field-Size Rate-Optimal Streaming Code for Channels with Burst and Random Erasures,” 2019 IEEE (5 pages). [cited by applicant]

Kuhn, N., et al. “Forward Erasure Correction (FEC) Coding and Congestion Control in Transport,” RFC: 9265, Internet Research Task Force (IRTF), Jul. 2022 (21 pages). [cited by applicant]

Le Gall, D., “MPEG: A Video Compression Standard for Multimedia Applications,” Digital Image and Video Standards; Communications of the ACM, vol. 34, No. 4, Apr. 1991 (13 pages). [cited by applicant]

Levine, S., et al., “Offline Reinforcement Learning: Tutorial, Review, and Perspectives on Open Problems,” arXiv:2005.01643v3 [cs.LG] Nov. 1, 2020 (43 pages). [cited by applicant]

Li, T., “Reparo: Loss-Resilient Generative Codec for Video Conferencing,” arXiv:2305.14135v1 [cs.NI] May 23, 2023 (17 pages). [cited by applicant]

Li, Y., “Deep Reinforcement Learning: An Overview,” arXiv:1701.07274v6 [cs.LG] Nov. 26, 2018 (85 pages). [cited by applicant]

Li, Z., et al., “Correcting Erasure Bursts with Minimum Decoding Delay,” https://www.comm.toronto.edu/˜akhisti/burst_erasure.pdf, pp. 33-39, 2011 (7 pages). [cited by applicant]

Li, Z., et al., “Forward Error Protection For Low-Delay Packet Video,” Proceedings of 2010 IEEE 18th International Packet Video Workshop, Hong Kong, Dec. 13-14, 2010 (8 pages). [cited by applicant]

Weiring, M., et al., “Reinforcement Learning, State of the Art,” Adaptation, Learning, And Optimization, vol. 12, 2012 (653 pages). [cited by applicant]

White, G., et al., “A Simulation Study of CoDel, SFQ-CoDel and PIE in DOCSIS 3.0 Network,” Active Queue Management Algorithms for Docsis 3.0, Cable Television Laboratories, Inc., Apr. 2013 (45 pages). [cited by applicant]

Wong, A., et al., “Deep multiagent reinforcement learning: challenges and directions,” Artificial Intelligence Review (2023) 56:5023-5056 (34 pages). [cited by applicant]

Wu, D., et al., “Transporting Real-Time Video over the Internet: Challenges and Approaches,” Proceedings of the IEEE, vol. 88, No. 12, Dec. 2000 (21 pages). [cited by applicant]

Yang, Y., et al., “An Overview of Multi-agent Reinforcement Learning from Game Theoretical Perspective,” arXiv:2011.00583v3 [cs.MA] Mar. 18, 2021 (129 pages). [cited by applicant]

Yaqoob A. et al., “A Survey on Adaptive 360 Video Streaming: Solutions, Challenges and Opportunities,” IEEE Communications Surveys & Tutorials, vol. 22, No. 4, Fourth Quarter 2020 (38 pages). [cited by applicant]

Zadnik, J., et al., “Image and Video Coding Techniques for Ultra-low Latency,” ACM Computing Surveys, vol. 54, No. 11s, Article 231. Publication date: Sep. 2022 (35 pages). [cited by applicant]

Zhang, K., et al., “Multi-Agent Reinforcement Learning: A Selective Overview of Theories and Algorithms,” arXiv:1911.10635v2 [cs.LG] Apr. 28, 2021 (73 pages). [cited by applicant]